If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

OMG, Google Swapped Your AI Image Model and You Didn't Notice

Did you generate an image in Gemini this week?

Then you used a different model than last week. And Google didn't exactly send a memo.

On February 26th, Google quietly replaced Nano Banana Pro (the model behind all those viral AI product images) with a brand new model called Nano Banana 2. No popup. No "hey, we changed things." Just... a swap.

If your images looked a little different and you couldn't figure out why, now you know.

Here's what actually changed, what it means for your product photography workflow, and how to switch back to Pro if you want to.

First, Let's Untangle the Version Soup

Google's naming makes this confusing. So let me lay it out clearly.

There have been three Nano Banana models. Each one is built on a different version of Google's Gemini AI. Here's the timeline:

Notice the pattern?

The original Nano Banana went viral in August. Built on Gemini 2.5 Flash, it was fast and fun, but rough around the edges. Then in November, Google released Nano Banana Pro on Gemini 3.0 Pro. Slower, more expensive, but with studio-level quality, better text rendering, and real-world knowledge.

That was the model people fell in love with. The one generating those jaw-dropping product shots.

So Did Google Downgrade Everyone?

Here's the part that has people confused. And a little frustrated.

On February 26th, Google replaced Nano Banana Pro (Gemini 3.0 Pro) with Nano Banana 2 (Gemini 3.1 Flash) as the default.

The version number went up, from 3.0 to 3.1. But the tier went down, from Pro to Flash.

It's not a version downgrade. It's a tier swap. Google moved everyone from the heavy, premium Pro architecture to the lighter, faster Flash architecture. The new Flash is better than the old Flash. But it's still Flash, not Pro.

Think of it like going from a BMW 3 Series to a newer Toyota Camry. The Camry is newer, more efficient, and honestly great for daily driving. But it's not a BMW.

The Silent Downgrade Nobody Asked For

And it gets worse for free users. Even before Nano Banana 2 launched, people on the Gemini community forums were reporting silent downgrades. When free users hit their daily image quota, Google would quietly drop them from Nano Banana Pro to the original Nano Banana. No notification, no warning.

Resolution would tank. Text rendering would degrade. And users had no idea why their images suddenly looked worse.

Now with Nano Banana 2 as the default, free users don't even get the option to use Pro. It's reserved for Google AI Pro and Ultra subscribers only.

Is Nano Banana Pro Going Away?

With all the changes happening, I had to check. Good news: Nano Banana Pro is not being shut down. Google's official deprecation page shows no sunset date for the image generation model.

So why the switcheroo? I think this is a cost play, not a quality play. Nano Banana Pro runs on the heavy Gemini 3.0 Pro architecture. It's expensive to operate at scale. Giving hundreds of millions of free users unlimited access to your most expensive model isn't sustainable. So Google did what makes sense: they moved the masses to the cheaper, faster Flash model and reserved Pro for paying customers.

For most casual users generating memes and profile pics? Nano Banana 2 is more than enough. But for eCommerce sellers who need that extra finesse (accurate product details, precise text on packaging, photorealistic lifestyle shots) Pro still matters. And now it's behind a paywall.

What Is Nano Banana 2, Really?

So where does this leave us? Three models, simplified:

🔶 Nano Banana (original). The OG. Gemini 2.5 Flash. Fast, fun, viral. But limited quality. Think of this as V1.

🔶 Nano Banana Pro. The premium upgrade. Gemini 3.0 Pro. Slower, pricier, but the best image quality Google offered. The craftsman.

🔶 Nano Banana 2. The new default. Gemini 3.1 Flash. Takes the best features from Pro (text rendering, world knowledge, sharper details) and runs them on the faster Flash architecture. The workhorse.

Google decided the workhorse was good enough for most people. So they made it the default and tucked Pro behind a menu.

The Numbers Don't Lie

Here's the side-by-side breakdown:

Half the price. 2-3x faster. 95% of the quality.

For most eCommerce workflows? That math is hard to argue with.

I Tested It. Here's What I See.

I ran the same simple product image through both models (a Focallure Face compact) to see the difference with my own eyes.

Make a model hold this product in her hand

Original Photo

👇 This is what I got with Nano Banana 2:

With Nano Banana 2

👇 And this is what I got with Pro:

Nano Banana Pro

The difference? Nano Banana 2 still has that telltale AI look. You know the one. An extra finger sneaking into frame, skin that's just a little too smooth. Pro? It looks like a real photo.

Pro gives you slightly cleaner edges, a touch more realism in the skin texture, better placement and a bit more polish overall. It's the difference between a really good photo and a really good photo

Nano Banana 2 is warm, clean, and production-ready. For a product listing? For a social media ad? For 90% of what we do as sellers? You'd never notice the gap.

But here's the thing. If you're doing high-end brand photography or need absolute pixel-level precision, Pro still has the edge.

How to Access Each Model

This is the part most people are missing.

Nano Banana 2 (Default. You're Already Using It)

Just generate an image in Gemini. That's it. Nano Banana 2 is now the default across:

🔶 Gemini App: select "Create images" from tools, works in Fast, Thinking, or Pro mode
🔶 Google AI Studio: via the Gemini API (paid key required)
🔶 Google Search: through AI Mode and Lens
🔶 Vertex AI: for enterprise deployment

Nano Banana Pro (Hidden, But Still There)

Here's the trick nobody's talking about:

1. Generate your image normally in Gemini (it uses Nano Banana 2)

2. Click the three-dot menu (⋮) on the generated image

3. Select "Regenerate with Nano Banana Pro"

That's it. Pro is one click away, but only if you know it's there.

Important: You need a Google AI Pro or Ultra subscription to access Nano Banana Pro. Free tier users only get Nano Banana 2.

The "So What" for Sellers

Let's do some quick math.

Say you're generating 50 product images a week for listings, A+ content, and social. That's a realistic number for an active seller.

🔶 With Pro: 50 images x $0.06 = $3.00/week, plus ~25 minutes of generation time
🔶 With Nano Banana 2: 50 images x $0.03 = $1.50/week, plus ~10 minutes of generation time

Over a month that's $6 in savings and an hour of your life back. Not earth-shattering for small volumes. But scale that to 500 images for a catalog refresh or a seasonal push? Now you're looking at $15 saved and 2+ hours recovered per batch.

Speed matters more than cost here. When you're iterating (trying 10 different angles, tweaking prompts, A/B testing lifestyle shots) a 10-second generation time versus 30 seconds changes how you work. You stay in flow. You experiment more. You get to "good enough" faster.

That's the real advantage.

The One Catch

Content filters.

Nano Banana 2 is noticeably stricter than Pro. Google tightened the safety rails around:

🔶 Famous figures and celebrities
🔶 Character face/outfit swapping
🔶 Anything remotely suggestive
🔶 Financial information in images

For most product photography, this won't matter. But if you're doing creative lifestyle content, brand storytelling, or anything that pushes boundaries, you might hit walls that Pro wouldn't throw up.

Also worth noting: character consistency across multiple images is still a weakness. If you need the same model holding your product across five different scenes, both models will struggle, but Pro handles it slightly better.

Pro or 2? My Take.

Here's how I'd think about it:

Use Nano Banana 2 (default) when:

• You're generating product-on-white or simple lifestyle images

• You're iterating quickly and need volume

• You're running A/B tests on creative

• Budget matters (it always does)

• You need readable text on packaging, labels, or signage

Switch to Pro when:

• You're creating hero images for brand storefronts

• You need the absolute best photorealism

• Content filters are blocking your creative direction

• You're doing a small batch where time doesn't matter but quality does

For 90% of what eCommerce sellers do every day? Nano Banana 2 is the move. It's fast, it's cheap, and the quality is close enough that your customers will never know the difference.

Pro is your secret weapon for the shots that really matter.

The Bigger Picture

Here's what I keep thinking about.

Six months ago, Nano Banana Pro was the cutting edge. People were blown away. It topped the image generation leaderboards.

Now it's the backup option behind a three-dot menu.

That's how fast this moves. The premium model of yesterday becomes the free tier of tomorrow. The tool you spent weeks mastering gets replaced by something faster and cheaper before you've even optimized your workflow.

This is why the learning window matters. Not because any single tool will last forever, but because the skill of adapting to new tools is the real advantage. The sellers who tested Pro last month are already comfortable with 2 this month. The ones who waited are still figuring out what changed.

Nano Banana 2 isn't perfect. Pro isn't dead, yet. But the direction is clear: faster, cheaper, and good enough is eating slower, expensive, and slightly better.

For eCommerce sellers, that's not a threat. That's an opportunity.

PPC Ninja is helping brands future proof their listings for AI, helping you build RUFUS enabled, stunning images and videos with AI. Hit reply on this to chat with us. Explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

Claude Remote Control 🤖 The Safe (But Limited) Alternative to OpenClaw

If you've been following the OpenClaw hype, you know the promise: an AI agent that runs on your computer 24/7, browses the web, manages your calendar, controls smart home devices, and basically acts like a digital assistant on steroids.

Sounds amazing. Until you hear that a recent security audit found over 820 malicious skills in OpenClaw's community hub. That's not a typo. Over 20% of available skills were flagged.

So when Anthropic dropped Claude Remote Control on February 25th, a lot of people (myself included) got excited. It's not the same thing as OpenClaw (more on that in a second), but it scratches a similar itch: controlling an AI agent from your phone. And it's way safer.

What It Actually Does

Claude Remote Control lets you run a Claude Code session on your computer and then control it from your phone, tablet, or any web browser. Think of it like screen-sharing, but for your AI coding assistant.

Here's the setup:

1. Open your terminal and run claude remote-control

2. A session URL appears. Press spacebar to show a QR code

3. Scan the QR code with the Claude app on your phone (or paste the URL in any browser)

4. You're now controlling your local Claude session from anywhere

Already inside a Claude Code session? Just type /rc to start Remote Control without leaving.

Want it on by default? Type /config inside Claude Code and set "Enable Remote Control for all sessions" to true. After that, every session you start is automatically available from your phone. No extra commands needed.

Pro Tip: Before going remote, type /rename to give your session a descriptive name (like "product page redesign"). It makes finding the right session on your phone way easier.

Your files, your tools, your project configuration. All still running locally on your machine. Nothing moves to the cloud. The phone is just a window into your local session.

Why It's Safer Than OpenClaw

🔶 No inbound ports. Your machine only makes outbound HTTPS requests. Nobody can connect toyour computer
🔶 Everything runs locally. Your code, your files, your data stay on your machine
🔶 No community skill marketplace full of unvetted (and potentially malicious) plugins
🔶 Short-lived credentials. Sessions are secured via TLS through Anthropic's API
🔶 You stay in control. It doesn't run unattended. You initiate, you guide, you approve

But Here's Where It Falls Short

Let's be honest about the limitations:

🔶 Not a 24/7 agent. Your terminal has to stay open. Computer goes to sleep? Session dies.
🔶 One session at a time. No running multiple remote sessions in parallel
🔶 10-minute timeout.Lose network for more than 10 minutes and you have to restart
🔶 No browser automation. It can't browse the web, check flights, or manage your email like OpenClaw can
🔶 No persistent memory.OpenClaw saves context to local files and can reference conversations from weeks ago. Remote Control starts fresh each session
🔶 Expensive. Requires a Max plan ($100-$200/month). The official docs say Pro support is "coming soon," but it's not there yet. Not available on Team or Enterprise plans either

So Who Is This For?

If you're a seller who also vibe codes (and I know a lot of you are), this is genuinely useful. Kick off a coding task on your laptop, then check on it from your phone while you're at the warehouse or on a call. Review code, approve changes, monitor progress.

It's not OpenClaw. It's not trying to be. OpenClaw wants to be your autonomous digital employee. Remote Control wants to be a really long extension cord for your existing Claude Code workflow.

My take: If you're already paying for Claude Max and using Claude Code, this is a no-brainer. If you're not in that world yet, this alone isn't worth the $100/month. But keep it on your radar. Anthropic is building toward something bigger and safer here.

Have you tried Remote Control or OpenClaw? I'm curious which camp you're in. Hit reply and let me know.Have you tried Remote Control or OpenClaw? I'm curious which camp you're in. Hit reply and let me know.

The Vertex AI Back Door to Nano Banana Pro 🤓

Now this section is for the Nerds. Remember how I mentioned above that Google moved Nano Banana Pro behind a paywall? That free users are stuck with Nano Banana 2, and Pro now requires a Google AI Pro or Ultra subscription?

Well, there's a back door. And it's sitting right there in plain sight.

Vertex AI: The Access Nobody Talks About

Most sellers interact with Nano Banana through the Gemini app. That's the front door. Limited daily generations, resolution caps, silent downgrades when you hit your quota.

But Vertex AI is Google's enterprise AI platform. And it gives you direct, unrestricted access to both Nano Banana Pro and Nano Banana 2. No three-dot menus. No praying your daily quota hasn't expired. Just the model you want, when you want it.

Here's the part that gets interesting:

🔶 Google Cloud gives new accounts $300 in free trial credits. At roughly $0.13 per 2K image, that's over 2,200 Nano Banana Pro images before you spend a dollar
🔶 Google AI Studio (a separate tool) gives free users 50 API requests per day with just a Google account. No credit card needed
🔶 On paid Vertex AI tiers, there are no daily caps. Generate as many images as your budget allows

Why This Matters for Sellers

While everyone else is fighting over 2-3 free images per day in the Gemini app, you could be generating hundreds through the API with trial credits. No subscription. No waitlists. Just direct access to the model you actually want.

This is especially powerful if you're doing catalog refreshes, seasonal creative pushes, or A/B testing across dozens of product variations. The Gemini app was built for casual users. Vertex AI was built for production.

How to Get Started

1. Sign up at cloud.google.com (you'll need a credit card for verification, but you won't be charged during the trial)

2. Enable the Vertex AI API in your Google Cloud console

3. You now have $300 in credits and full access to Nano Banana Pro

That's it. Three steps and you've bypassed every limitation the free tier throws at you.

Pro Tip: If you're not comfortable with APIs yet, start with Google AI Studio (aistudio.google.com). It gives you a visual interface to test prompts with 50 free requests per day. No coding required. Think of it as the middle ground between the Gemini app and full Vertex AI.

Go With The Flow Podcast: Cowork Deep Dive - Skills, Guardrails, and Avoiding Disaster

In this episode Danny McMillan and I dive deep into Anthropic's new Cowork feature in Claude Desktop. The good, the bad, and the ugly.

I share a cautionary tale about file deletion gone wrong (yes, it happened to me), and Danny demos his custom skills pack that protects users from common pitfalls.

What we cover:

🔶 What Cowork is and how it differs from Claude Code and Claude Desktop
🔶 Why rm -rf can permanently delete your files (and how to prevent it)
🔶 How to set up deny lists in your Claude settings to protect critical files
🔶 The power of skills and bootstrap files for consistent, reliable outputs
🔶 2026 predictions: Ambient AI and pixel-free interfaces

Have a listen!

Convert More Clicks Summit
RSVP here: https://maximizingecommerce.com/CMC26-Ritu-Java
It’s FREE!

Super excited to be speaking at the Billion Dollar Seller Summit BDSS 13, and AI-focused event hosted by Kevin King. RSVP: https://www.bdss13.com

🚀 Never Pay for Another App! Learn How to Build Custom Features Yourself📍 Event: Seller Summit 2026
📅 Dates: April 21–23, 2026
Save your spot here: Link

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

What is Prompt Injection?

You've probably heard the term floating around. Prompt injection. It sounds technical. It sounds like a developer problem.

It's not. If you're using AI in your business (and you should be), this is something you need to understand.

The simplest way to think about it:

Prompt injection is when someone hides instructions inside content that your AI reads, and the AI follows those hidden instructions instead of yours.

Remember writing secret messages with orange juice as a kid? The paper looked completely blank. But hold it over heat and the hidden message appeared.

Prompt injection works the same way. Attackers hide instructions inside normal-looking content (product reviews, emails, documents) that are invisible to you but get "revealed" when an AI reads them.

The AI doesn't know the difference between your instructions and theirs. It just follows what it reads.

I built a little Prompt Injection Detector simulation to show you what I mean. It works like holding invisible ink over a flame.

In the left pane you see a perfectly normal product review. Click the reveal button and the right side lights up with hidden instructions that were lurking inside the whole time.

That entire red block was hiding in the review, sandwiched between invisible Unicode characters. You couldn't see it. Your customer couldn't see it. But an AI reading that review? It sees everything, and it follows the instructions without question.

Here’s a example of a potential prompt injected message from a supplier. See what happens when we do the flame test:

That's bad enough with one AI. Now picture a whole team of them.

The industry is moving fast toward multi-agent systems. Multiple AI agents working together autonomously. One handles your ad bids, another writes your listing copy, another manages your customer service queue. The potential for e-commerce sellers is massive.

But prompt injection doesn't just scale with multi-agents. It compounds. Here are three risks worth knowing:

🔶 The Domino Problem. In a multi-agent setup, agents hand off tasks to each other. If one agent picks up a poisoned instruction (say, from a tampered product data feed), it doesn't stop there. It passes the bad instruction down the chain. Your inventory agent feeds bad data to your pricing agent, which feeds bad data to your ad agent. One contaminated input, the whole workflow is compromised.

🔶 The Backdoor Problem. Your agents don't all have the same power. One might only read reports. Another might send emails. Another might authorize payments. But if they share a common memory layer (and most multi-agent frameworks do), then compromising the weakest agent can give an attacker indirect influence over the most powerful one. It's like leaving your warehouse unlocked because "it's just storage," until someone uses it to access the office safe.

🔶 The Multiplier Problem. With a single AI, one prompt injection affects one system. With multi-agents, a single malicious instruction buried in an email or document can get picked up and re-executed by every agent in your network. What used to be a one-off trick becomes an automated chain reaction.

The good news? The big players are taking this seriously.

Sam Altman just brought the creator of OpenClaw, Peter Steinberger, into OpenAI specifically to tackle multi-agent coordination and security. That's a strong signal that these problems aren't being ignored.

But you don't need to wait for OpenAI to protect you. Some practical guardrails you can put in place today:

🔶 Treat every external input as untrusted. Product reviews, supplier emails, customer messages. Anything an AI agent reads from the outside world should be filtered before it enters your workflow.

🔶 Lock down agent permissions. Every agent should have access to only what it needs. Your analytics agent shouldn't be able to write to your payment system. Period.

🔶 Keep a human in the loop for high-stakes actions. Automated bidding? Fine. Automated refunds above a threshold? That needs a human approval step.

🔶 Audit the handoffs. The riskiest points in a multi-agent system are where one agent passes information to another. Know where those handoffs are. Log them. Review them.

The sellers who build with guardrails from day one won't need to wait for someone else to protect them. The learning window is open. Understanding this risk early is the competitive advantage.

Claude Code's Playground: Build Interactive Tools in Seconds

A new badass functionality in Claude opened up. It’s called “Plugins” and you can literally add on hundreds of Plugins that come with certain pre-trained skills. Today I am talking about a Plugin called “Playground”

To set it up, just access Plugins in the bottom left…

Then Browse available plugins…

Look for Playground and click Install:

Now you can just access this plugin from your Claude Code chat bar like so…

The Playground skill lets Claude generate self-contained interactive tools directly in your browser. Sliders, presets, live previews, all running locally in a single HTML file. No frameworks, no dependencies, no deployment. You describe what you want, Claude builds it, and it opens in your browser ready to use. Like this one I built (feel free to click and check it out):

That's an Amazon PPC Dashboard playground. KPI cards, budget alerts, hourly heatmaps, chart layouts. All configurable with the controls on the left, all updating in real time on the right. Built with one prompt. Now, remember that this is a playground, and each time you make an adjustment it will translate into a prompt that gets fed right back into Claude.

Five ways people are using Playground:

🔶 Visual dashboards. Turn your campaign data into interactive HTML reports you can actually explore instead of staring at spreadsheets.

🔶 UI prototyping. Need a landing page layout? Describe it, get a live preview with sliders for spacing, colors, and typography. Tweak until it's right, then send it back to Claude to build the real thing.

🔶 Product image prompt builders. Create a playground with controls for lighting, angle, background, and mood. See your prompt update live, copy it, paste it into your image generator.

🔶 Document review tools. Load any text file and get inline suggestions you can approve, reject, or comment on. Like a visual editor for your listing copy.

🔶 Data exploration. Got a complex search term report? Turn it into an interactive page where you can filter, sort, and spot patterns visually.

Here are Five prompts to try right now with Playground:

🔶 PPC Dashboard

Create a playground for an Amazon PPC Dashboard. Controls: 
date range selector, KPI card picker (ACoS, ROAS, impressions, clicks, CTR, CPC, CVR), chart layout toggle. Preview shows KPI cards with sparklines, budget alerts, ad spend vs sales chart,and hourly performance heatmap. Dark theme.

🔶 Product Landing Page Builder

Create a playground for a product landing page layout. Controls: hero image position (left, right, full-width), feature columns(2, 3, 4), color scheme presets (warm, cool, neutral), CTA button style, font pairing selector. Preview shows a live product page for a premium kitchen appliance brand.

🔶 AI Product Photo Prompt Builder

Create a playground for building AI product photography prompts. Controls: product type (bottle, box, pouch, jar), surface material, lighting style, camera angle, mood. Preview shows the assembled prompt updating live. Include presets: Clean Minimal, Luxury Editorial, Lifestyle Warm.

🔶 Listing Copy Reviewer

Create a playground for reviewing Amazon listing copy. Load a sample listing with title, bullets, and description. Show inline AI suggestions with approve, reject, comment buttons. Include readability score and keyword density meter in the sidebar.

🔶 Search Term Report Explorer

Create a playground for exploring Amazon search term data. Sample data with 50 terms including impressions, clicks, spend, sales, ACoS. Controls: sort by column, filter by ACoS threshold, spend range slider. Highlight high-spend zero-sale terms in red. Scatter plot of spend vs sales.

PPC Ninja is helping brands future proof their listings for AI, helping you build RUFUS enabled, stunning images and videos with AI. Hit reply on this to chat with us. Explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably. 

The OpenClaw Ecosystem

Geo from one of my Whatssap groups on AI recently shared this cool OpenClaw ecosystem and I thought it was worth passing along. Thanks Geo!

It’s crazy to think that less than a month ago, OpenClaw did not even exist!

Want to learn the science behind RUFUS?

Andrew Bell put together a neat resource called Generative Artificial Intelligence Model Streaming: An interactive educational guide to Amazon's patent on efficient AI response streaming with retrieval-augmented generation and intermediate token processing.

As you might now, Andrew eats patents for Breakfast, Lunch and Dinner. So you can expect this to be a solid resource if you want to satisfy your intellectual curiosities.

This is an interactive resource that let’s you learn at your own pace. You can even attempt a quiz at the end to test your understanding. Here’s the link to check it out.

When Kevin King puts up an event, you know it’s going to be stellar. I am SUPER excited to be speaking at the ECOM Mastery AI featuring Billion Dollar Sellers Summit in Nashville. When: April 8-12, Nashville. Get your tickets here.

🚀 Never Pay for Another App! Learn How to Build Custom Features Yourself📍 Event: Seller Summit 2026
📅 Dates: April 21–23, 2026
Save your spot here: Link

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

🗞️ AI is now Writing Itself

For the last two years, we treated AI like a very fast intern. Give it a task. Get a draft. Move on. Output was the end of the interaction.That model is already obsolete.

When machines started operating on their own instructions, something fundamental shifted. Anthropic recently said that Claude is now writing Claude code. Claude Cowork was written entirely by Claude.

When an agent distills documentation into a compact Skill and feeds it back into itself, we are no longer using AI. We are managing recursive systems.

This is the Recursive Feedback Loop. Output is no longer an endpoint. It becomes structured input for the next run. Every task leaves the system sharper than before.

From Context Rot to Context Compaction

For a while, we believed more context meant better results. Bigger prompts (pages and pages of it). More PDFs. Longer threads.

What we got instead was context rot. The model knew more, but understood less. Signal buried in noise.

The breakthrough was not bigger context windows. It was Context Compaction.

Instead of feeding chaos into models, operators force the model to distill what matters. The result is dense, reusable logic rather than raw information.

This is not summarization. It is compression with intent.

What “Skills” Actually Are

When Claude introduced the concept of “Skills” the idea was simple. Skills are the unit of compaction. As of today, Anthropic is not the only AI that supports skills. Gemini and Codex support it too. In practice, they are simple markdown files. Plain text. High density. Deliberate structure.

They live outside the model. In repos. In folders. In version control. They are loaded when needed, not rediscovered through prompting.

That is why Skills persist across sessions, agents, and models. They survive context limits and prompt churn. They allow intelligence to accumulate.

As a Claude user, you can add Skills under Capabilities in your Settings. Claude can even help you create Skills out of work that you are already doing.

If you work at the command line or with Claude Code, the folder structure can contain a Claude.md file that Claude will automatically look inside for any instructions you left it. No need to even say “Go to Claude.md” or restate context every time.

What can e-Commerce Operators Extract from this?

Every serious e-commerce team already has Skills. They just live in people’s heads, Slack threads, and forgotten Notion docs.

Start pulling them out.

Turn your best PPC audits into Skills. One markdown file that encodes how you typically diagnose wasted spend, when you pause, when you scale, and what signals you trust. And then turn that into a Skill.

Turn your brand voice into a Skill. One markdown file that encodes how your brand speaks, what it prioritizes, which claims it allows, which comparisons it avoids, and what “off brand” actually means. Load it once, and let every listing, ad, image, and response inherit it by default.

Turn image generation into a Skill. Encode your visual rules, composition preferences, lighting, backgrounds, do’s and don’ts, and what a “good” image looks like for your brand. Load it once, and let every creative inherit consistency instead of starting from scratch.

Store these Skills where your tools already look. Project folders. Repos. Campaign directories. Let your agents inherit them automatically instead of re-explaining them every time.

The goal is not more automation or better prompting. It is fewer resets.

If your AI needs you to re-explain how you run your business every week, you are still operating in 2024. The teams winning in 2026 are the ones turning experience into compact, persistent memory.

What Skills are you creating with AI? Hit reply and let us know!

Google Flow

A notable shift at Super Bowl LX is how AI became both the subject and the engine of ads. About 23 percent of commercials (15 of 66) referenced or featured AI in some way, with tech players and consumer brands alike pushing the narrative that AI was now part of mainstream life. That includes giants like OpenAI, Anthropic, Google Gemini, Amazon, Meta, Wix, and others buying premium Super Bowl airtime to frame their products through artificial intelligence messaging.

The most AI-heavy commercial of the night was Svedka’s “Shake Your Bots Off”. The vodka brand claimed it was the first primarily AI-generated Super Bowl spot, featuring its robot mascots animated using generative tools. That ad leaned fully into the technology as a creative statement even if reactions were mixed.

The Svedka spot mattered not because it was brilliant, but because it crossed a line. AI was no longer hiding behind polish or pretending to be invisible. It was the point. The tool became the headline. And once that happens on the Super Bowl stage, the question stops being whether brands should use AI and starts being whether they know how.

Of the many film making tools out there, Google Flow is a far choice that is aimed far above quick clips. This is text to cinema. You describe scenes, characters, tone, and pacing, and Flow generates short narrative films using Google’s Veo models. The output feels directed, not stitched together. Camera movement is deliberate. Lighting holds. Characters stay expressive instead of melting between frames.

Most Flow videos today live in the short film range. Think several seconds per scene, often extended into one to two minute narratives by chaining scenes together. The quality jump is what stands out. This is no longer novelty AI video. Depth, shadows, facial emotion, and cinematic framing hold up on large screens. You can still tell it is generated, but it no longer looks disposable.

One of the standout features is Ingredients. Ingredients let you define and reuse the building blocks of a story. Characters, locations, props, visual style, mood, lighting, even camera behavior can be saved and reused across projects. This solves the biggest pain in generative video which is drift. Your character stays your character. Your world stays intact. Your brand look does not reset every time you prompt.

Ingredients unlock something most product videos never touch. Story continuity. A character that appears across launches. A customer avatar who returns in different moments. A consistent visual universe that feels intentional instead of promotional.

Flow is not built for feature demos. It is built for narrative. Origin stories. But with some creativity and a bit of Canva magic, you can convert it into a decent video that can be displayed in your storefront.

PPC Ninja is helping brands future proof their listings for AI, helping you build RUFUS-enabled, stunning images and videos with AI. Hit reply on this to chat with us. Explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

Gemini Gems

I have always been a fan of Gemini Gems. They are closest thing Google has to CustomGPTs, but built to live where the work already happens. Over time, I have moved many of my CustomGPTs over to Gems, mainly because they stay inside the Google ecosystem instead of sitting beside it.

You will find a ton of premade Gems by Google that are ready to use out of the box. Or, you could create your own from scratch with your own special instructions to perform certain tasks consistently.

Gems are not floating chat personas. They can reference Google Drive files, work inside Workspace apps via Gemini, and increasingly plug into Notebook LM as a structured source of truth. Instead of pasting docs, copying links, or reminding the model how to behave, you design a Gem once and let it operate on curated knowledge. When that knowledge changes, you update the source, not the prompt.

One smart workflow I saw recently, shared by John Aspinall, pairs Gems with Notebook LM. Notebook LM holds the long term memory, things like PPC frameworks, CTR research, DSP notes, brand strategy docs. The Gem handles execution, turning that material into SOPs, drafts, or structured documents when needed.

Notebook LM becomes the library.
Gems become the operator.
Google handles the wiring.

When Kevin King puts up an event, you know it’s going to be stellar. I am SUPER excited to be speaking at the ECOM Mastery AI featuring Billion Dollar Sellers Summit in Nashville. When: April 8-12, Nashville. Get your tickets here.

🚀 Never Pay for Another App! Learn How to Build Custom Features Yourself📍 Event: Seller Summit 2026
📅 Dates: April 21–23, 2026

Save your spot here: Link

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

New here? Welcome. Been around a while? Thanks for sticking with me through this AI shift in eCommerce. Each week: what I'm experimenting with, what's working, and trends worth catching early.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

The OpenClaw Takeover: This is What Happens When AI Moves Faster Than You Can React

In under a week, an obscure AI agent went through 2 name changes, triggered a Mac Mini buying frenzy, jumped from bedrooms to the cloud, and inspired what might be the first social network built by bots, for bots.

Here’s your time lapse to this fascinating piece of history:

Day 1 (Circa Jan 26, 2026): An open-source agent called ClawdBot crosses an invisible line and goes viral. It starts spreading rapidly across Reddit, not as a demo, but as a working system. This isn’t a chatbot. It can see screens, click, browse, and act autonomously. Early adopters stop testing it and start handing it real work.

Day 2: The Mac Mini and Mac Studio craze. Builders start buying dedicated machines just to run ClawdBot nonstop. Desks turn into mini data centers. This is no longer “tinkering”. Things are heating up.

Day 3: The lawyers arrive. ClawdBot is renamed Moltbot almost overnight because of the similarity to “Claude”. The community shrugs and keeps shipping Mac Minis.

Day 4: With so many security vulnerabilities, suddenly running agents on Mac Minis feels dumb (ha!) Everyone moves to VPS setups so agents can live in the cloud instead of under desks. And in parallel, the Moltbots have already built their own social media platform, a variant of Reddit cutely named Moltbook.

Day 5: A new religion for bots is formed and fierce discussion among bots ensues around whether a new “agents-only” language is required.

Day 6: Cloudflare (respectable, publicly traded company) drops Moltworker, demonstrating how Moltbot can run securely on its platform using sandboxes, browser automation, memory, and access controls. DIY gives way to production. Around the same time, Moltbot settles on its third name: OpenClaw.

Three names. Five days. No slowdown.

By the time you finish reading this, some of it will already be old news.

The Million Dollar Question: Is it Safe?

Short answer: it depends on how you use it.

What just happened in five days wasn’t about better AI. It was about architectures maturing in public. Local machines gave way to VPS. VPS gave way to Cloudflare-grade isolation. The direction is obvious: autonomy is getting safer, cheaper, and easier to deploy.

That doesn’t mean “hand over the keys.”

It means the game has changed.

From automation to agency (why this actually matters)

Most sellers are familiar with automation. Rules-based tools. If this, then that. Safe, predictable, limited.

Agency is different.

An agent like OpenClaw doesn’t just follow instructions. It decides. It observes context, takes action, checks results, and adapts. If a checkout breaks, it doesn’t log an error. It finds another path. If customer support backs up, it doesn’t alert a human. It restructures the workflow.

That’s why people are uneasy. And they should be.

Not because the tech is reckless, but because agency removes friction faster than humans are used to.

What should E-Commerce Sellers do next?

If you’re an e-commerce seller, especially on Amazon or Shopify, the opportunity here is not “replace the team.” It’s compress feedback loops. But first you need to learn what’s even possible.

Mike Michelini shared this about how he is putting his bots to use as an Agent Swarm Fleet, where one bot manages the work of others:

Common and emerging ways OpenClaw is getting used right now:

• Personal assistant tasks: email triage, calendar scheduling, reminders, and inbox clean-up.

• Messaging automation: controlling the agent via WhatsApp, Telegram, Discord, iMessage, etc.

• Browser automation: filling forms, navigating websites, extracting data.

• Shell and system commands: executing scripts, managing files, launching local processes.

• Multi-agent workflows: separate agents handling email, schedule, and coordination between them.

• Reusable skill modules: connectors for productivity, finance, home automation, and development workflows.

• Persistent memory: remembering preferences, context, and past tasks to improve future decisions.

• Crypto/finance experiments: autonomous tracking of markets, yield strategies or on-chain monitoring (reported in niche communities).

• Self-organizing agent interactions: agents engaging each other on Moltbook, trading info and patterns.

• Bot-to-bot social dynamics: threads, upvoting, shared conventions, and emergent coordination among AI agents.

The Moltbook lesson most people are missing

Moltbook isn’t interesting because bots talked about religion. It’s interesting because they coordinated without us.

Once agents interact at scale, they stop being isolated helpers and start becoming systems. They influence each other. They reinforce patterns. They optimize in directions humans didn’t explicitly design.

That’s powerful. And dangerous if you’re not watching.

For sellers, the takeaway is simple: never run agents you can’t observe.

Logs matter. Explainability matters. Kill switches matter. Human-readable checkpoints are not optional.

The real advantage sellers can build right now

The sellers who win this phase won’t be the ones who automate everything first.

They’ll be the ones who are actively learning. They:
• Spend time with their bot to work on its own security. Things like closing open ports, securing firewalls, disallowing root login, removing password login
• Learn how agents behave under pressure and understand where they fail silently
• Design constraints before scaling autonomy
• Treat agents like junior operators, not gods

This is a learning window. It won’t last long.

Soon, these capabilities will be embedded into tools you already use or other viral open source replacements. When that happens, the advantage won’t be access. It will be experience.

Final thought

This isn’t hype. And it isn’t a warning. It’s a reminder that software no longer waits for instructions. The OpenClaw moment isn’t about bots replacing humans. It’s about humans learning how to supervise systems that move faster than they do. The sellers who get this right won’t chase speed blindly.

They’ll move fast, yes. But with guardrails.

One last laugh to close it out:

Credit to Monte Desai for sharing this viral post

🎯 My Hack from the 10x5 Amazon Hacks Webinar: Turn Plain English into Full Production Workflows with Google Opal

Tired of hearing about vibe coding but feeling like you need a technical background to get started? I shared a hack that removes that barrier entirely.

Introducing Google Opal – a free tool that lets you build AI-powered workflows using nothing but natural language. No code. No complex setup. Just describe what you want, and Opal builds it for you.

I demo'd a viral social post generator I built in Opal that:

• Takes a topic + product photo + brand logo

• Researches viral content angles

• Generates the image and copy

• Outputs a ready-to-post social media asset

Think of it as Lego blocks: User Input → AI Generate → Output – and you just talk to connect them.

📺 Watch the full walkthrough in the video below to see exactly how I built it.

PPC Ninja is helping brands future proof their listings for AI, helping you build RUFUS enabled, stunning images and videos with AI. Hit reply on this to chat with us. Explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

Go with the Flow Podcast: Cowork Deep Dive - Skills, Guardrails, and Avoiding Disaster

What happens when you give AI full access to your computer?

In this episode, Danny and Ritu explore Anthropic's new Cowork feature, a powerful tool that lets Claude autonomously handle complex tasks on your desktop. They dig into what it can do, where it shines, and the hard lessons learned when things go sideways.

Plus: Danny shares his custom skills pack that makes Cowork actually reliable, and they wrap up with some bold predictions for where AI agents are headed in 2026.

If you're thinking about letting AI do more of the heavy lifting, this one's worth your time. Have a listen..

podcasts.apple.com/kg/podcast/go-with-the-flow-cowork-deep-dive-skills/id1304966378?i=1000747399349

Super excited to be speaking at the ECOM Mastery AI featuring Billion Dollar Sellers Summit in Nashville. When: April 8-12, Nashville. Get your tickets here.

🚀 Never Pay for Another App! Learn How to Build Custom Features Yourself

📍 Event: Seller Summit 2026
📅 Dates: April 21–23, 2026
 Save your spot here: Link

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

The Era of Disposable Software

For decades, we built digital monuments. Overbuilt platforms designed to last years but slow to adapt and costly to change.

But when the cost of code trends toward zero, the value of durability starts to disappear. In this new paradigm, what matters is effectiveness today, not perfection over time.

This is Vibe Coding, an approach where you don't need to identify as a developer. You just need to know what's possible. And now you can DIY your way into apps that improve productivity, create images, videos, market your products, write email, the list goes on..

One message rang loud and clear at BDSS 13. The power to build AI software at scale has shifted from engineers to visionaries.

What was once the domain of backend teams, architecture reviews, and multi-month sprints can now be executed by a single person with a clear idea and an AI co-pilot. And that shift is about more than just convenience. It's about rethinking the very nature of software itself.

We have entered The Era of Disposable Software.

From Vision to Execution in Minutes

In my talk at BDSS 13, I shared how this shift is playing out in my own business.

Here are some examples of Vibe Coded software:

• I recently replaced the Hubspot Sales CRM with a custom-built Google Sheets app powered by Claude and deployed with CLASP. Did you know that you can easily build software apps and add them as Menu items to any Google Sheet so that can manipulate your business data?

• I recently replaced Wix, redesigned the ppcninja.com site from the ground up using Claude Code. The site files live in an AWS S3 bucket and CloudFront points to that bucket. This cut my hosting and editing costs from $1,000 per year to $10 per year.

• At PPC Ninja we have built over 40 AI-powered tools without writing traditional code. One of them even does Bid + Placement optimization for us weekly. Crazy times we live in!

Vibe Coding is not about being a hacker or having deep technical knowledge. It's about seeing what's possible and being willing to build, test, and discard quickly.

🗑️ Software Built to Be Deleted

In this new era, the best tools are:

• Fast to build

• Inexpensive or free

• Easy to throw away without regret

Code is now just a napkin sketch. You capture an idea, test it quickly, and ship it. If it works, great. If not, move on.

And this is where Git becomes so important. It acts as your undo button for reality, giving you the freedom to experiment without risking your core business.

📁 Files Are the New Memory

One of the most important ideas I shared is this: AI forgets. Files remember.

Your chat history does not hold context, but your local project files do. Markdown files like:

• CLAUDE.md

• STYLEGUIDE.md

• BUSINESS_LOGIC.md

act as persistent memory for your AI assistant. They provide structure, continuity, and business logic across every project.

⚙️ AI as Interface Instead of Engineer

You do not need to learn the command line or the nitty gritties of how to deploy code. That belongs to an older era. In this new world, AI is your interface. It translates your intent into commands using tools like:

• AWS CLI for hosting

• FFMPEG for media manipulation

• CLASP for Google Apps scripting

You describe what you want, and the AI gets it done.

PPC Ninja is helping brands future proof their listings for AI, helping you build RUFUS enabled, stunning images and videos with AI. Hit reply on this to chat with us. Explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

Claude Cowork (The Agent in Your Machine)

Ok, everyone is talking about Claude Cowork. The tool that was built by Claude Code in 10 days and released. Wow!

We are moving from AI that talks to AI that does. Claude Cowork is the first mainstream attempt at an "agent", an AI that can remote control your desktop, click buttons, and execute complex plans. It doesn't just write code; it runs it. It doesn't just suggest organization; it moves the files.

After a week of testing, the results are interesting: it is brilliant at local file manipulation and maybe not so great yet at navigating the open web or Google Drive. Here is exactly what it can do for you right now, and where it fails.

The Wins (What You Can Do Today):

• The "Downloads" Detox: Point it at your chaotic Downloads folder. Cowork can analyze thousands of files, create a taxonomy (e.g., "Invoices," "Installers," "Personal"), and physically move them into new folders. WORD OF CAUTION (Read my full story below…)

• Visual File Renaming: This is the killer feature. You can ask it to "Rename all these screenshots based on what they show." It uses vision capabilities to see the image and rename Screenshot_29.png to AI_News_Anchor_Collage.png.

I asked Cowork to clean up the last 2 months of screenshots by reading the images and giving it appropriate names. I also allowed it to create sub-folders for better organization.

To do this you will first need to point it at a local drive folder.

Point to the folder and grant it permissions to “clean it up”. MAKE SURE YOU RESTRICT HOW FAR BACK TO GO. I asked it to only go back 2 months.

Here’s the “safe prompt I used”

Cowork acknowledged the instruction and proceeded to rename and reorganize my folders.

I thought it took a bit long (almost like 10 minutes, which for us these days is looooong)

Here are some other things you can do with Claude Cowork:

• Format Alchemy: It excels at transforming data types. You can dump a folder of disparate benchmark screenshots and ask for a unified Excel spreadsheet. It can read a PDF and generate a video outline, then convert that outline back into a PDF.

• Local Media Operations: It can run code locally (like FFmpeg). You can ask it to find a specific video clip ("Find the video with the yellow leaves") and then convert that clip into a GIF.

• The "Meta" Report: If a task fails (and it will), you can ask it to generate a report on why it failed. It will analyze its own error logs and produce a polished Google Doc explaining the process.

The Fails (Where to Hold Off):

• Web Navigation: It struggles with the "human" internet. In tests, asking it to upload files to Google Drive failed because it couldn't physically click the right buttons due to browser restrictions.

• The "Continue" Button: It is not fully autonomous. On long tasks (like renaming 1,000 images), it gets stuck and requires you to nudge it by typing "continue" repeatedly.

The Actionable Strategy:

To get value out of Cowork, you must change your workflow. Do not ask it to "browse the web and do X." Instead, use the "Container Strategy."

1. Curate the Input: Manually gather the messy files (transcripts, screenshots, PDFs) and put them into a single folder.

2. Grant Access: Give Cowork access to that specific folder only.

3. Prompt for Synthesis: Ask it to perform the work inside that folder. "Read these 5 transcripts and write a summary report," or "Turn these research notes into a slide deck."

The Verdict:

Claude Cowork removes the friction between "messy data" and "structured asset." It is not a replacement for a human assistant yet, but for digital janitorial work and data synthesis, it is the best tool on the market.

But I haven’t finished my horror story yet, proceed to the next section…

How to Avoid the Horrors of Losing all your Data!

As a next step I thought, well, Claude Cowork did well, I may as well have asked Claude Code to do the folder reorganization. Perhaps it will be bit faster and more intelligent than Cowork.

So I pointed it at the same folder and asked it to do the same thing - reorganize my files from 2 months ago, which were now all sitting in the Recent (Nov-Jan 2026) folder.

Claude Code was moving fast, and I was like wow, it’s doing it! It even told me how it was renaming all my files based on a better understanding of what each screenshot was. Nice.

And that’s when things suddenly went wrong! Claude Code accidentally deleted all those files that it just finished renaming.

You see, Claude used the rm -rf command to bypass the Trash folder. Ha, the story that others on the internet were talking about and I thought I’d be immune to!!! Thankfully the damage was very minimal, since this was just a subfolder in my Screenshots folder.

After a bit of digging I found that there is a way to prevent Claude from ever removing file permanently (i.e. bypassing the Trash folder)

For all my nerdy peeps, this is what you’ve got to do right away:
Edit your Global Claude settings ~/.claude/settings.local.json

"deny": [
      "Bash(rm *)",
      "Bash(*&& rm *)",
      "Bash(*; rm *)",
      "Bash(*| rm *)"
      "Bash(rm -rf *)",
      "Bash(rm -fr *)",
      "Bash(rm -r *)",
      "Bash(rm -Rf *)",
      "Bash(*&& rm -rf *)",
      "Bash(*&& rm -r *)",
      "Bash(*; rm -rf *)",
      "Bash(*; rm -r *)"
    ]

The -rf argument is a skip without asking operation and by adding it to the Deny list you can prevent accidental deletion of entire folders while still allowing normal single-file deletion.

How to set up this Deny list? Give the above to Claude and ask it to do it for you. Stay safe!

ChatGPT is Getting a Shopping Cart!

You’ve probably heard it already, but shoppers will be able to purchase products directly from ChatGPT. OpenAI is rumored to charge 4% of the sale price.

As Max Sinclair points out, “If your product can’t be purchased natively, it will be skipped altogether. Now’s the time to:

• Adopt the agentic commerce protocol

• Make sure your product data is clean and structured in accordance with OpenAI's specs

• Apply for instant checkout via OpenAI or enable it directly through Shopify

Jan 27th is the last Tuesday of the month. Time to tune in to the Go with the Flow Podcast where Danny McMillan and I dive into AI topics. Listen here!

Join me for this incredible Hack Contest on Jan 28th organized by Smart Scout and Nick Penev. 17 hacks in 90 minutes!! RSVP here.

Super excited to be speaking at the ECOM Mastery AI featuring Billion Dollar Sellers Summit in Nashville. When: April 8-12, Nashville. Get your tickets here.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

Ambient AI and the Rise of the Invisible Storefront

For twenty years, ecommerce obsessed over the eye. Better photos. Cleaner layouts. Faster loads. A quiet belief that if it looked good enough, someone would buy. We built systems designed to impress humans who were willing to browse, compare, and decide for themselves.

We spent decades polishing the surface of ecommerce, perfecting images, copy, and layouts built to dazzle. But the signal coming out of CES and the NRF Big Show was unmistakable. Intelligence is leaving the interface and moving into the plumbing.

One idea kept resurfacing in different forms. We are in the Year of the AI Agent. Digital commerce already accounts for more than half of retail transactions, and now, in a growing number of cases, the first customer is no longer human. It is an agent acting on someone’s behalf. On Amazon, that agent is Amazon Rufus. It is Alexa Plus.

Your storefront is no longer designed for browsing. It is designed for final selection. By the time a shopper lands on your page, the real evaluation has often already happened somewhere else.

At NRF 2026, Shopify and Google introduced the Universal Commerce Protocol, an open standard designed to let AI agents connect to merchants and complete transactions autonomously. That detail matters because it removes the last assumption that a human must always be present at checkout. The loop is closing because Ambient AI is here.

The invisible storefront is the reality where buying decisions are made by AI agents upstream, and your brand only exists if it can be clearly understood, compared, and chosen before a human ever sees the page.

AMBIENT AI

From destination to integration

We are moving away from destination commerce, where shopping meant going somewhere and searching. We are entering ambient commerce, where buying happens where you already are. Voice assistants, predictive prompts, and background decisions replace deliberate journeys.

The search bar fades into the background. Typing becomes optional. Intent is inferred from context, history, and constraints. As a result, SEO quietly gives way to Answer Engine Optimization, not ranking pages, but supplying certainty.

This is the Zero UI world. Commerce becomes a trigger, not a trip.

ALEXA PLUS

The machine to machine economy

This is the deeper shift underneath all of it. Humans are no longer the primary gatekeepers. Their agents are. We are entering a machine to machine economy where software evaluates options, compares tradeoffs, and makes recommendations before a person even looks.

If your product cannot be clearly understood by an agent, it effectively does not exist.

A useful analogy surfaced at NRF. Tools built for people with visual impairments translate the visual web into structured meaning so it can be navigated confidently. Your storefront now has to perform the same translation for AI agents. If the bot cannot see your value through data, you are closed.

How to build an invisible storefront

The idea is not to look better, it is about being clearer.

🔶 Attribute density
Every backend field matters. Materials, dimensions, weight, use cases, constraints. Agents filter on specifics, not vibes.

🔶 Semantic narratives
Bullet points should answer objections, not decorate features. Rufus matches intent, not adjectives. Benefits beat polish.

🔶 Conversational Q and A
Your FAQ is no longer defensive copy. It is training data. Comparisons, tradeoffs, and pros and cons are what agents summarize out loud.

The quiet advantage

Most brands will respond to Ambient AI by producing more content. That instinct is right. But in an agent-driven market, content only works when it increases certainty. Unstructured volume slows selection. Structured depth speeds it up.

Your website does not disappear, but its role changes. It becomes a verification layer, where humans go only when the agent needs confirmation. The real competition happens earlier, inside the structured truth an AI uses to classify, compare, and justify a recommendation.

The brands that win will not be the ones with less content or louder messaging. They will be the ones whose content removes doubt fastest. Easy to explain. Easy to compare. Easy to choose.

PPC Ninja is helping brands future proof their listings for AI, helping you build RUFUS enabled, stunning images and videos with AI. Hit reply on this to chat with us. Explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

Deploy Google AppScript by Chatting with AI

I create a lot of Google AppScripts for all kinds of use cases. You’ve heard me talk about my mini apps living inside the Google ecosystem. Recently, I ran across a SUPER handy method that cut my build time even further.

The magic ingredient is Clasp. It’s a small command line tool from Google that lets scripts be created, updated, and deployed without opening the Script Editor.

No copying. No pasting. No spreadsheet gymnastics. On a Mac, I install it through Homebrew. On Windows, it’s the same idea without Brew. Install once, sign in once, and your computer is trusted to manage Apps Script projects for you.

brew install node
npm install -g @google/clasp

AUTHENTICATE ONCE: 
clasp login

Here’s where it gets interesting. I’ve hooked this up to Claude Code. That means I can talk through what I want, Claude writes the Apps Script files, and then actually runs the command line steps to push and deploy the script using clasp. I’m not just getting code back. I’m getting working software dropped straight into my Google account.

The shift is subtle but huge. Apps Script stops being something you “implement” and starts being something you describe. If you can explain a workflow, you can build it. Clasp just handles the boring logistics in the background while AI does the heavy lifting.

The Backlash from Shopify Owners

Last week, Amazon got caught building agentic features on top of independent ecommerce websites without their permission. In the context of what everyone else in the industry is working on, that felt out of place. Amazon is hacking together a prototype Buy for Me function while also fighting every major AI lab that wants to train on its data. Meanwhile, the rest of the ecosystem is doing the unglamorous work of building shared foundations. Protocols. Integrations. Agreements. Like the recent Universal Commerce Protocol backed by Google and two dozen partners.

Here is a viral video that talks about what happened:

Instagram post

The backlash from Shopify sellers is a tell. When Buy for Me started scraping independent stores and replaying them inside Amazon, the issue wasn’t legality. It was intent. Invisible storefronts still have owners. Agents still act on someone’s behalf. Pretending that layer is neutral is how trust breaks quietly.

Juozas Kaziukėnas wrote an insightful post highlighting this. Read the full piece:

Super excited to be speaking at the Billion Dollar Seller Summit BDSS 13, and AI-focused event hosted by Kevin King. RSVP: https://www.bdss13.com

Join me for this incredible Hack Contest on Jan 28th organized by Smart Scout and Nick Penev. RSVP here.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

I Fired Wix

For 5 years, ppcninja.com ran on Wix, a popular hosted website builder.

It did its job, and it was ok. But as the business matured, the friction became harder to ignore. The editor was clunky. Responsive behavior was unpredictable. And every small change carried overhead. Find someone on Upwork, explain what should have taken minutes, pay $100 to $200, wait, and move on.

That model works at a certain stage. It just didn’t scale with how we operate now.

What pushed the decision wasn’t dissatisfaction. It was volume. We needed a lot of changes made, quickly. We wanted samples of the creative work we were doing uploaded fast, and we wanted to update blogposts on PPC topics each week.

And it was obvious the same process would repeat itself again. More updates = More coordination = More time spent managing the tool instead of improving the site.

So what had been background friction turned into a one day experiment. By the end of it, ppcninja.com was running a gorgeous, fully responsive static HTML website, hosted on AWS-S3, fronted by CloudFront, and Wix was gone.

Annual savings: about $700.
Time spent: one focused day.
Trade offs: fewer than you think.

Here’s what changed.

Instead of logging into a visual editor and nudging boxes around, I described the changes in plain English. A short conversation with Claude. What I wanted to change. How I wanted it to look. A few sentences at a time. Claude translated that into clean static HTML that I could review, tweak, and iterate on immediately.

Once the HTML was ready, the hosting part was almost boring. Claude showed me exactly how to do it. Upload to S3. Add a CloudFront distribution. Point the domain. Done. The whole thing felt less like “web development” and more like configuring infrastructure, which is exactly the point.

Wix sells convenience. And for a long time, that trade off made sense. But the moment AI can generate usable frontend code on demand, that value equation changes. Fast.

What surprised me most wasn’t the cost savings. It was the mental shift. The website stopped feeling like a product we rented and started feeling like an asset we owned.

No monthly platform tax.
No feature gating.
No fear of breaking things by clicking the wrong setting.

This is where the superpower actually shows up.

Claude didn’t just write copy. It didn’t just “help.” It replaced an entire category of SaaS. Not because it’s better at hosting, but because it collapsed the distance between idea and implementation.

Vibe coding with Claude

That’s the pattern to pay attention to.

AI doesn’t always replace tools directly. Sometimes it removes the reason the tool existed in the first place.

I’m not suggesting everyone should rip out their CMS tomorrow. But if you’re paying hundreds every year for something that boils down to static pages, it’s worth asking a different question now.

Think hard. What else are you renting that AI can help you own?

This is what I mean when I say superpowers don’t arrive as flashy features. They show up quietly, shave real costs, and give you back control.

Weavy AI

Weavy is a visual workflow canvas for building repeatable AI powered creative processes.

At its core, Weavy is a visual workflow canvas for creative AI. You build flows by connecting blocks, run the same prompt through multiple models, branch outputs, and layer in real editing steps like cropping, masking, resizing, compositing, or pulling frames from video. It feels less like prompting and more like building a repeatable machine.

The signal that this category is real: Figma acquired Weavy. That’s not a casual pickup. It’s a strong bet that visual, node based AI workflows are becoming part of mainstream design tooling, not just a playground for power users.

What makes Weavy especially fun is the model buffet. In a single canvas, you can route prompts through multiple image models side by side. Flux variants, Imagen, GPT Image, Higgsfield, Reeve, and more. Different aesthetics, different strengths, same input. Instead of debating which model is “best,” you just let them all compete.

On top of generation, Weavy includes a surprisingly deep toolbox. Levels, painter, crop, resize, blur, channels, mask extraction, mask by text, even video frame extraction. This is the stuff that usually forces you into Photoshop or After Effects. Here, it’s just another node in the flow.

The real unlock is reuse. Once you build a workflow you like, it becomes a template. You are no longer reinventing your process every time. You are refining a system.

If you care about AI generated visuals at any kind of scale, Weavy is worth a look. Not because it’s flashy, but because it turns experimentation into infrastructure.

How much you ask? There is a free plan with 150 credits followed by $19 a month for 1500 credits. Definitely worth a test!

PPC Ninja is helping brands future proof their listings for AI, helping you build stunning images and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

5 ways to move faster with AI

1. Double-tap the Control key to talk to your AI
On Mac, double-pressing the Control key triggers voice input, and it’s fully configurable in system settings. If you are on Windows, use a software like Willow Voice or any AI launcher, and you can start talking to your AI instantly. No typing. No context switching.

The real trick is not trying to sound polished. Speak messy. Half thoughts. Corrections mid sentence. AI is very good at cleaning up ideas. It’s terrible at inventing the ones you never said.

2. Use special-purpose markdown files as “first read” context
Files like AGENTS.md, PROJECT.md, or README.md are increasingly treated as first-stop context by modern LLM tools, especially in coding and repo-based workflows.

Use them to define what the model should never guess:

• Project goals

• Hard constraints

• Style preferences

• Tools or libraries to favor or avoid

Keep them short and opinionated. Think onboarding note, not documentation.

3. One project, one context file
If you work across multiple initiatives, create a small markdown file per project with:

• Objective

• Audience

• Non-negotiables

• Success criteria

Load it once at the start or let your tool auto-attach it. This prevents drift without forcing long prompts every time.

4. Tell the model what not to do
Negative constraints are underrated.

Examples:

• “Do not explain basics.”

• “Do not add marketing fluff.”

• “Do not introduce new tools.”

This reduces exploration cost and keeps the model inside useful boundaries.

5. Snapshot good work before switching models
Before you jump from one model to another, pause and extract a clean snapshot:

• Final prompt

• Key assumptions

• Best output so far

Paste that into the next model as starting context.

This avoids both extremes: overly long conversations and cold starts. You carry forward only what earned its place.

Super excited to be speaking at the Billion Dollar Seller Summit BDSS 13, and AI-focused event hosted by Kevin King. RSVP: https://www.bdss13.com

Join me for this incredible Hack Contest on Jan 28th organized by Smart Scout and Nick Penev. RSVP here.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here. Also, this is the last post for 2025! Hope you have enjoyed our journey together so far :)

GPT 1.5 or Nano Banana? Who is the New Visual King

Just when we thought Google was winning the AI image creation race, OpenAI threw a systemic curveball that shifted the battle from raw pixels to cognitive control. While Google’s Nano Banana Pro optimized for the "perfect shot" through high-resolution fidelity and reference-image anchoring, GPT Image 1.5 doubled down on the logic of the frame. It treats complex spatial requests like a physics engine rather than a guessing game. It didn't just provide a better brush; it provided a smarter art director that understands the semantic relationship between objects, effectively ending the era of "pretty but random" generations in favor of total architectural intent.

The Comparison Challenge

To see this shift in action, I ran a "Stress Test" prompt designed to break typical AI, testing complex lighting, specific text, and anatomical precision. You can run this yourself to see the divergence:

This is what I got back from Nano Banana:

Nano Banana

And here is GPT 1.5…

GPT 1.5 (Image)

The Breakdown: Surgical Precision vs. Studio Power

GPT Image 1.5 is the "Design Partner." In professional work, having to "start over" is a massive waste of time. GPT 1.5’s superpower is Instruction Persistence.

Unlike Nano Banana, where the canvas is largely static, GPT 1.5 allows you to edit specific sections of the screen while keeping the rest of the image pixel-locked. You can also queue up multiple images while you wait, refining them with surgical precision. It behaves like a high-end retoucher, making it the superior choice for AI-Ready Listings. It generates a structured asset where the text, lighting, and product placement stay exactly where you put them.

Furthermore, it respects the professional pace: you can queue up multiple images while you wait, refining them with surgical precision.

Nano Banana Pro is the "Studio Powerhouse." Powered by Gemini 3, it remains the king of material realism. While GPT 1.5 is the better architect, Nano Banana is the better "photographer." It pushes native 4K resolution and captures the "grit" of reality—micro-scratches on glass, natural skin pores, and complex lighting physics—without the slightly glossy "AI sheen." For high-end print or "hero" shots where the subsurface scattering of light on a product must be indistinguishable from a Leica photograph, Google holds the technical edge in raw atmospheric depth.

The Split into Elements Test

Let’s do another comparison. Ask GPT 1.5 and Nano Banana to split an image into elements with a transparent background. I took this image:

Original Photo

GPT 1.5 correctly split all the elements out. Amazing.

GPT 1.5 did it right

Unfortunately, Nano Banana Pro did not get the assignment.

Nano Banana did not split all the elements correctly

Extracting Labels with Ease

If you have one clear shot of your product with clear labels, with GPT 1.5 Image, you can now extract the label and transfer the label on let’s say, a blurry product shot, like I did:

Product with Clear Label

Product in hand blurry label

Final Image: Clear Label Replaces Blurry Label

Sweet. See how clean and clear the new product image is?

Which One Should You Use?

Too many choices. That’s the reality of the AI era we live in.

• Choose GPT Image 1.5 if: You are building AI-Ready Listings at scale. If you need to generate 100 variations of a product in different room settings while maintaining 100% brand consistency, this is your engine. Its ability to "lock" the product and iterate on specific environment sectors is the new industry standard for e-commerce and real estate.

• Choose Nano Banana Pro if: You need gallery-grade realism. If your final output is a 60-inch print, a luxury brand hero image, or a shot that requires authentic location details, Google’s 4K pipeline is the gold standard.

The Verdict: A Professional Blend

At our PPC Ninja creative studio, the choice isn't binary; it’s a blend. We use both models as a tag team to cover the entire production spectrum. However, after producing hundreds of professional listing images for real-world clients, my team’s preference has become clear: they are liking GPT 1.5 a lot better. 

The reason? Predictability. When you’re managing high-volume listings, you can’t afford an AI that "hallucinates" a new layout every time you ask for a minor tweak. GPT 1.5’s ability to lock in the product while we iterate on the environment saves us hours of prompt-engineering. It feels less like a game of chance and more like a high-performance workstation.

PPC Ninja is helping brands future proof their listings for AI, helping you build stunning images and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

Gemini is Now Available in Chrome!

Google has reimagined the browser as an "agentic" tool, moving beyond simple searches to a system that can autonomously perform multi-step tasks—like planning a trip or cross-referencing data across ten tabs—on your behalf. However, these features are not active by default; they require specific steps to unlock.

Step 1: Go to this page and download the all-new AI enabled Chrome browser.

Step 2: Install it and refresh your Chrome and start using it.

Step 3: Enter AI Mode in the Omnibox

The address bar (Omnibox) is now the primary entry point for agentic browsing. To move beyond a standard Google search and into the reasoning engine:

• The Shortcut: Type your query into the address bar. Look for the AI Mode sparkle icon that appears on the right side of the bar.

• Action: Click the icon or hit Tab + Enter. This shifts Chrome into "Thinking" mode, where it can navigate websites, read content, and fill out forms to complete your task.

Other cool things you can do with this new and improved Omnibar.

Start typing @ to access a bunch of shortcuts

Your new browser can now interactively answer any questions about a page. To activate it, simply click into the omnibar and select “Ask about the page”

Google Lens lets you select any section of any page and now you can chat with Gemini about it.

Have you tried it yet? Hit reply to this email and share your thoughts.

The Power Duo: GPT + Canva

The integration of GPT and Canva transforms a text-based AI into a visual strategist. By invoking the @Canva command directly within your chat, you bypass the friction of starting with a blank canvas. Instead of scouring templates manually, you describe your vision, and the assistant bridges the gap between your concept and professional design layouts.

Activate Canva by going to your GPT Settings → Personalization → Apps → Explore Apps and then connect to your Canva app.

Once you gave activated Canva and hooked it up in your account, you can simply activate it by saying @Canva at the GPT command line and then type your prompt.

When I prompted Canva to generate a product poster, it didn't just offer a single static image; it produced four distinct design directions. This variety is crucial for A/B testing or simply seeing how different visual hierarchies—like bold typography versus image-heavy layouts—impact your brand’s message.

The real utility lies in the seamless handoff. Each of the four generated options remains fully dynamic. Clicking a design opens it directly in Canva, where every element—from hex codes and font weights to grid spacing—is editable. You aren't stuck with an AI’s "best guess"; you have a high-fidelity starting point that you can refine to match your exact brand guidelines.

This workflow effectively eliminates "blank page syndrome" for non-designers. By combining GPT’s linguistic precision with Canva’s drag-and-drop flexibility, you can move from a rough product concept to a polished marketing asset in minutes. It’s no longer about "generating an image," but about collaborative iteration between AI speed and human curation.

Why This Combo Wins:

• Speed to Market: Reduces the design cycle from hours to seconds.

• Infinite Customization: Moves beyond static AI images into editable vector/layered formats.

• Contextual Relevance: GPT understands the intent of your poster, ensuring the initial templates align with your industry or tone.

In his latest update on RUFUS, Andrew Bell highlights a pivotal shift from passive AI to agentic action. He pointed out that that the real power of Rufus lies in its "Actions Stack", a sophisticated loop of Memory, Reasoning, and Action that moves beyond simple chat to execute complex shopping missions.

The most standout capability is the ability to step outside the Amazon eco-system and provide intelligence. Read the full article here.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

The Code Behind the Canvas: Why JSON is the Secret to AI Consistency

Most people think "style" is about adjectives. They try to replicate a look by typing “make it moody, cinematic, dark, blue lighting.” The result? Usually a caricature. It captures the vibe but misses the soul.

I recently discovered a better way, and it changes everything about how we use AI for creative work. It’s not about finding better words; it’s about changing the language entirely. I started feeding AI images and asking it to reverse-engineer them into JSON code.

What in the world is JSON?

JSON (JavaScript Object Notation) is a code format that acts as a blueprint, locking in specific variables like lighting and texture so the AI doesn't have to guess. This is what JSON looks like.

Wait, what? Code to describe an image?

That’s right. It sounds counterintuitive: using code to describe beauty. But structured data is the native language of AI. It cuts through the ambiguity of English.

The "Forensic" Workflow

I started with a static image. A high-gloss cosmetic ad featuring jellyfish by Japanese beauty brand, Kate. Instead of describing it, I asked the AI to analyze it and output the style as a JSON object. This gave me a digital fingerprint of that aesthetic: specific parameters for lighting, color grading, and composition, stripped of the subject matter.

When I fed that JSON back into the LLM, the results were striking. I could produce dozens of images that looked exactly like they came from the same campaign. Just look at this…!

It worked so well that I built a custom Gemini Gem just to handle this "Style-to-JSON" workflow.

Why JSON Mimicry Works (The "Slot Machine" Theory)

Why does this work better than normal prompting?

When you write a paragraph of text, you are forcing the AI to guess the relationships between words. "Dark" might modify "lighting," or it might modify "mood." The AI has to interpret your grammar.

JSON removes the guesswork. It forces the AI to decouple the Subject from the Style. Think of it like a slot machine.

• Text Prompt: You pull the handle, and all the wheels spin together. If you change the subject, the lighting often changes with it accidentally.

• JSON Prompt: The "Lighting" and "Camera" slots are locked. You only spin the "Subject" wheel.

By speaking to the model in data rather than prose, we bypass the "hallucination" of style. We aren't asking the AI to imagine a look; we are giving it the blueprints to build it.

From Images to Video

I didn't stop at static images. Once I realized JSON acts as a "rendering engine" rather than a chatbot, I applied it to video.

Steve Simonson recently shared a JSON for an Apple video. I took that JSON code, plugged it in into VEO3, and the AI recreated that distinct, high-fidelity Apple "feel" instantly. No long paragraphs of description, just structured data.

Just look at this beauty…

Have you tried using JSON for AI image style mimicry? How did the results turn out for you?

Breaking the Glass: The "Bulk File Automator"

The biggest frustration with AI is amnesia. You build a perfect workflow in a chat, but when you close the window, it’s gone. If you want to analyze an Amazon Ads bulk file, you usually have to paste messy snippets of data and explain the column headers every single time.

There is a better way: Claude Code. You can access in the top left of your Claude app window in your browser (no need to go to the desktop version.)

This tool allows the AI to break out of the browser and access a specific folder on your actual desktop. It transforms Claude from a chatbot into an operator that can read, edit, and generate files locally on your laptop.

The Workflow

I created a folder on my desktop called Bulk File Automator. I activated Claude Code in this directory and gave it one instruction:

Create or update my CLAUDE.MD file with all the knowledge necessary to create and edit Amazon bulk files, documented below. The user will then use this project to work on various bulk files in the future.

In the directory is a TEMPLATE Amazon Bulk File with the right header rows for each ad type, and a SAMPLE of a file with data.

The documentation from Amazon is here: https://advertising.amazon.com/API/docs/en-us/bulksheets/2-0/overview-about-bulksheets

Now, that folder is intelligent!

It "knows" the Amazon schema permanently. I can drop my actual bulk file (CSV/XLSX) into that folder and execute three powerful use cases:

1. Chat with your Bulk File You don't need to explain the columns—it already knows them.

"Are there any campaigns with an ACoS > 50% that are set to Fixed Bids?"

2. Extract & Create Subsets

"Extract the top 10 performing keywords and save them as a new file called top_performers.csv."

3. Create New Files on the Fly You can ask it to generate complex upload files from scratch.

"Create a new bulk file for a campaign named 'Jellyfish_Launch'. Set the daily budget to $50, default bid to $1.50, change strategy to 'Down Only', and include these 5 keywords."

Because you "trained" the folder with the documentation URL, it generates a perfectly formatted file ready for upload—no manual formatting required.

You know that feeling when you upload a bulk file and it uploads error free :)

Amazon’s AI shopping assistant has launched Deep Research Mode, a capability that moves Rufus from simple Q&A to a comprehensive research agent. By running up to 30 searches and synthesizing data from external editorial sources, Rufus now builds personalized, "Wirecutter-style" buying guides in seconds.

Why this matters for brands: Rufus is no longer selecting products based solely on sales ranking. It now prioritizes role, context, and use-case fit. To win in this new environment, brands must shift focus toward "Reasoning Path Optimization" (RPO) and external authority building to ensure they aren't just ranked high, but are understood as the right solution for the shopper’s specific goal.

Read the full post by Andrew Bell on LinkedIn.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

You are Paying Hallucination Tax

You know that lovely message:
“You have reached the end of your chat limit. Conversation too long”

That is not your LLM showing you attitude. That’s your context window hitting concrete.

Roughly speaking:

🔶 Gemini 1.5 Pro gives you a context window up to around two million tokens in dev land. That is hours of audio, long video, entire codebases in one go.

🔶 GPT 5.1 gives you a very large window in the API and a generous window inside ChatGPT on the Thinking tier so you can work with serious documents instead of tiny snippets.

🔶 Claude Opus 4.5 sits on the Claude long context family around the two hundred thousand token mark for most users with special long context modes that stretch higher.

On paper it sounds like one simple story. Bigger windows. Fewer broken chats.
In practice, each lab is making a different trade.

Claude, Gemini, OpenAI and where the burden lands

All major LLMs are juggling the same three pieces

1. Context window what the model can see right now

2. Compression what gets summarized or compacted when that fills

3. Retrieval or memory how old info gets pulled back in

Gemini 3.0 leans on retrieval. It indexes huge piles of content then pulls the right slices into that big window when you ask a question. The headline is long context. The real trick is which pages it chooses to show the model.

OpenAI GPT 5.1 pushes on size and structure. Big windows in the API and a separate memory layer in ChatGPT for stable preferences to save your preferences and instructions.

Anthropic Opus 4.5 looks fussy because it ends long chats earlier. That is by design. Claude would rather stop than squeeze too much history into the window and start guessing. It seems counter-intuitive, but they are doing the right thing.

Opus 4.5 now compresses older turns so agents can run longer, but when answers might get blurry it still hits pause. So the user stays responsible for managing long context.

Gemini and OpenAI might seem to absorb more of the pain for you, but may silently be switching to hallucination mode without you realizing.

Anthropic protects answer quality and hands more control and more work back to the human.

Where the hallucination tax shows up

Every time a model compresses your history it keeps the headline and trims the fine print. The main story survives. The nuance does not.

Stack enough compression and you start paying hallucination tax

🔶 Numbers get rounded into friendlier stories
🔶 Edge cases and caveats disappear
🔶 The model confidently repeats something that was true three summaries ago and no longer is

Huge windows have their own failure mode. Treat Gemini or GPT 5.1 like an infinite dumping ground and retrieval starts to blur at the edges. The answer still sounds sharp. It is just quietly wrong in the third decimal place or the small print.

Bigger windows and smart compression stop chats from crashing.
They do not remove your responsibility to keep facts straight.

How to stop overpaying

If you are using these models for real work, treat context like budget.

🔶 Work in episodes
New project, new thread. Start with a tight recap. Update that one recap instead of expecting the model to reconstruct six weeks of side quests.

🔶 Re anchor in source when it matters
When money, contracts, or live data are involved, put the real spec or sheet or doc back in front of the model. Ask for answers that point back to that source, not to the chat history.

🔶 Do not chain summaries
If you need a fresh summary, ask the model to pull from the original doc or a rich outline, not from yesterday recap. A summary of a summary is where subtle hallucinations breed.

🔶 Split profile from project
Let long term memory handle who you are, tone, and reusable frameworks. Keep live numbers, SKUs, constraints and timelines inside the current context so they cannot be quietly warped by old compression.

🔶 Ask the model to show its doubts
Tell it to flag places where it might be guessing because earlier parts of the chat were compacted away and to ask you to restate facts. You trade a little magic for a lot more reliability.

Gemini long context, GPT 5.1 expanded windows, and Opus 4.5 careful compaction are real upgrades.

The edge goes to the people who still behave as if the window is small and act like editors of the context, not just consumers of the output.

Notebook LM for Infographic Style Explanations

You know that feeling when you finally finish a long, carefully crafter how-to guide and then realize most normal human beings will find it hard to follow?

That was my RUFUS extractor guide.

It walked through two separate JavaScript bookmarklets, how to turn them into browser bookmarks, what each one does, and the exact sequence for expanding RUFUS questions, scraping Q and A, and grabbing everything to your clipboard.

So I tried something different.

NotebookLM by Google has a cool bunch of tools to take any collection of documents, photos, website links or YouTube videos and turn them into any of these:

I asked Notebook LM to create an infographic style explanation for my RUFUS extractor and basically threw my entire document at it. You are looking at the result.

It separated the workflow into two clean parts, gave each its own color band, pulled out the exact three steps for the Clicker, the exact three steps for the Extractor, and turned my doom scroll instructions into a one screen, “oh, I get it” visual.

No designer. No extra prompts. Just “here is my guide, make it visual and simple.” I like it.

Now, could you have done this with Nano Banana? Yes, of course, but you would also have to direct the look and the prompt. More work.

With Notebook LM, none of that is required. It is like an intelligent analyzer of documents that was created with the the sole purpose of helping humans learn better.

AI video is not “someday” anymore

At UnBoxed, someone stopped by our booth, looked at our screens, and asked this question:

“So when do you think AI will be able to make realistic videos”

I did not hedge.
I said, “Now.”

Not “in five years.”
Not “once compute catches up.”
Now.

(GIF Loop of a small clip from a larger video we created)

Right now you can:

🔶 Take a single product image and turn it into a fully animated video clip
🔶 Generate dozens of short hooks in different styles for Amazon, TikTok
🔶 Iterate on motion, camera angle, and pacing with text prompts instead of reshoots

Is it perfect?
No.

Will you toss 9 out of 10 videos because of weird hands, odd physics, and the occasional uncanniness?
Yes!

But it is already good enough to extract segments that are fine, stitch them together with B-roll that already exists, throw in stills where you can, create 15 second Sponsored Brands video ads.

You’d rather be ahead of the race than wait for things to be perfect.

The future of e-commerce requires scale and speed. PPC Ninja helps brands dominate the AI transition. We leverage AI to build stunning, high-converting images and video, efficiently scaling your content production across all channels (Amazon Ads, Social Media, Posts). Reach out to support@ppcninja.com to explore how we can immediately upgrade your content and future-proof your listings.

I jammed on this topic with Danny McMillan in a recent conversation about AI video and what it means for brands that sell physical products. If you want a deeper dive, the full chat is here:

What we kept circling back to is this

The bottleneck is not the tech anymore.
The bottleneck is imagination, taste, and workflow.

If you are still waiting for AI video to reach some mythical “broadcast ready” milestone before you experiment, you are already behind the people who are using it today to

🔶 Prototype concepts before they pay for a studio
🔶 Test four story angles in a week instead of one per quarter
🔶 Give their creative team starting points instead of blank timelines

AI video is not here to replace your best work.
It is here to remove all the excuses between your idea and your first testable draft.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

If you’re new here, welcome! If you’ve been reading for a while, thank you for sticking around as we navigate this wild AI shaped shift happening across eCommerce. Each week I share what I’m experimenting with, what’s actually moving the needle, and the trends that deserve your attention before they hit your competitors’ playbooks.

A quick heads up. I’ve organized all previous editions into one searchable hub. If you want the full journey, it’s all here.

Visual reasoning finally grows up with Nano Banana 2

I have been testing the new Nano Banana 2 and it feels like someone slipped a spatial reasoning engine into a text model. It does not try to impress you with fireworks. It just sees. And once it sees, the things you can make it do become a whole new playground for ecommerce content.

Here are the five moments that made me sit up. Literally.

1. Builds spatial context

Prompt: Create a top view of the pink blush

BEFORE

AFTER

You give it an angled shot of a product and it understands how that object would look if you hovered over it. Not a guess. A structurally sound reconstruction. If you are creating Amazon content, this is a cheat code for angles you never shot.

2. Understands object composition

Prompt: Disassemble this toy into components of head, body, wheels, battery and remote

It does not cut things randomly. It identifies each part and lays them out like a product engineer prepping for an assembly guide. If you have ever fought with freelancers to make clean component breakdowns, this model just solved that.

3. Follows on-image guidance

Prompt: replace the marked area by a woman's hand holding this pink blush

BEFORE

AFTER

4. Adds cinematic details

Prompt: make this image into a masterpiece with cinematic details

BEFORE

AFTER

Most models either overcook or flatten. Nano Banana 2 does something more interesting. It studies the scene, figures out the vibe, and then lifts it into something that feels directed rather than generated.

5. Translates context into any language

Prompt: Turn the Japanese on image into English

BEFORE

AFTER

It reads the text, understands where it sits in the layout, and rewrites it naturally without breaking the design. For anyone selling globally, this is the fastest route to localized lifestyle images you have ever had.

Nano Banana 2 feels like an early signal of where image models are going. Less chaos. More awareness. A sense of space and intent that finally makes on image reasoning practical for day to day creative work. Have you tried it yet? Hit reply and share your favorite use case!

The future of e-commerce requires scale and speed. PPC Ninja helps brands dominate the AI transition. We leverage AI to build stunning, high-converting images and video, efficiently scaling your content production across all channels (Amazon Ads, Social Media, Posts). Stop fighting for relevance. Reach out to support@ppcninja.com to explore how we can immediately upgrade your content and future-proof your listings.

Hugging Face Image Angles

Nano Banana 2 can generate new angles on its own, but Qwen plus Kling gives you something different. Qwen lets you steer the camera with exact controls. Rotate by precise degrees. Push the lens forward in clean increments. Shift the vertical angle without guessing. You get a matched pair of shots that feel like they came from a controlled photo setup.

Step 1

Drop your image into the Hugging Face Qwen Image Edit tool. The magic here is the sliders. You can rotate the product by precise degrees, move closer in controlled steps, or shift the bird to worm angle without guessing. This gives you a clean pair of frames that feel like they came from a planned shoot.

Step 2

Send both frames to Kling AI. Add a simple prompt or run it through the DeepSeek enhancer inside Kling if you want a stronger arc. Kling reads the two angles and builds a smooth transition between them.

Watch the final clip. It looks like a motion product demo even though everything came from one photo and two sliders.

This is the power of combining tools! I find myself doing a LOT of that. In fact any AI Video generation projects we are taking on right now are a combination of 5+ tools to achieve one small effect to move it along in the pipeline. A lot of the video assets created are tossed. Almost 9 out of 10 are slop. So while AI is getting better and better, it is not easy!

The future of e-commerce requires scale and speed. PPC Ninja helps brands dominate the AI transition. We leverage AI to build stunning, high-converting images and video, efficiently scaling your content production across all channels (Amazon Ads, Social Media, Posts). Stop fighting for relevance. Reach out to support@ppcninja.com to explore how we can immediately upgrade your content and future-proof your listings.

Gemini 3.0 is out!

Gemini 3.0 is getting strong early reactions for one simple reason. It feels capable in places where most models wobble. Long tasks, technical steps, mixed language prompts. It handles all of it with a level of discipline people were not expecting.

You can activate Gemini 3.0 by simply selecting “Thinking”. The model takes more time per query and produces answers that reflect multi step thinking rather than quick pattern matching.

🔶 It handles multi part prompts with more discipline
Users give it long chains of instructions or nested conditions and it works through them step by step. No skipping. No forgetting earlier parts.

🔶 It explains reasoning without sounding fabricated
The Thinking mode shows its work in a way that feels closer to methodical problem solving. Developers say the explanations track the logic rather than decorating it.

🔶 It manages branching tasks better
People are testing it with prompts that require if this then that logic. The model follows the branches cleanly and keeps the structure intact.

🔶 It reduces shallow answers
The Fast mode jumps to conclusions. Thinking mode pauses and considers alternatives. The tone feels slower but more careful.

🔶 It improves coding and debugging quality
Programmers say Thinking mode is less likely to gloss over a broken function. It reads logs with more precision and offers fixes that match the actual issue.

🔶 It works well for planning and decision trees
Users give it business workflows, research plans, multi step checklists and it organizes them with fewer contradictions.

The best part is that you can pair it with Nano Banana and take things to the next level!

The overall vibe is simple. Gemini 3.0 behaves like a model built for real work rather than demo tricks. It is stable, structured and surprisingly reliable in long form reasoning.

Every 4th Tuesday of the month, Danny McMillan and I go live on the Go with the Flow podcast at 8 am PST on Seller Sessions! We dive into crazy nerdy things we’ve been working on related to AI. We discuss what’s real, what’s useful, and what’s next. Join us!

If you missed our last episode, we talked about Scraping in great detail. It’s a good one.

SellerVisor is an AI listing strategist for Amazon. Upload one product photo and an ASIN, and it will instantly analyze your top competitors to surface keyword gaps, image strategy ideas, and data-backed creative direction. All designed to help you convert better. New users get 14 days of full access, including 20 free image generation credits

→ [Try it free now]

With their AI-powered ASIN Comparison Tool, sellers can clearly see how their listing stacks up against the competition, image by image, keyword by keyword, and feature by feature. For professional teams, the pricing is $479/year:

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

Most people hear MCP (Model Context Protocol) and their eyes do the slow fade. It sounds like something reserved for people who sleep with YAML under their pillow. But the truth is MCP is quietly becoming the new plumbing of AI workflows. If you want your models to do anything useful with your real data, you’re going to bump into it sooner than you think.

MCP and e-Commerce

Amazon just rolled out their own Ads MCP server in beta, which tells you exactly where this is heading. Platforms are racing to give AI direct access to the real world. Amazon’s version slots into that trend. It’s their bridge between AI agents and the massive set of tools AWS already runs.

Early signs point to a model where a Claude or a GPT can call real AWS functions, hit real APIs, and actually perform small actions on your behalf inside your cloud stack. It’s still early, but the direction is clear. No more toy demos. Real data. Real actions. Real consequences.

Claude and BigQuery MCP Server

I recently wired up Claude desktop with Google BigQuery through the MCP server called MCP Toolbox.

Picture a tiny translator sitting between Claude and your database (in my case, BigQuery). You ask the model something. MCP handles the messy bits. Suddenly Claude is reading columns, shaping queries, and helping you reason about your data without you manually spelunking through SQL.

How to set this up?

I have to admit, it wasn’t an easy setup. There were folders, keys, configs, and a few moments where I wondered if I should just go back to spreadsheets. But once it clicked, the whole thing felt obvious. AI agents are not going to thrive in isolation. They need hooks into your systems. MCP is that hook. And you can ask AI to help you with the step-by-step.

To what end?

The fun part is what happens after the MCP server is actually connected. You stop “running SQL queries” and start chatting with your data. Simple things like asking what the average is across all clients at my agency. Or the harder stuff like why did my TACoS go up last month and letting the model dig through multiple tables and assemble the answer on its own.

A year ago this would have required real engineering. Custom scripts, scheduled jobs, layers of glue code. Now you point your model at an MCP server and it starts behaving like an internal analyst who already knows where everything lives.

That is the shift. You move from writing SQL to asking better questions. And the AI handles the messy cross table detective work that used to eat half a day.

What next?

If you’ve been watching this space from the shoreline, this is the moment to wade in. MCP is not hype. It’s the protocol layer that lets your models work inside your business instead of floating above it.

You don’t need to be an engineer to get the big picture. You just need to understand that the future of AI is not bigger models. It’s connective tissue. And MCP is quietly becoming the connective tissue of everything.

PixVerse: Pic to Video in no time

Video is still the heavyweight in e-commerce, but most brands get stuck at the same choke point. Shoots are slow. Edits are slower. Budgets vanish fast. PixVerse cuts straight through that friction. It turns a single product photo or a short text prompt into multiple motion clips you can actually test. Think Image to Video, Text to Video, and those viral effects that social feeds love. One shot becomes five variations in minutes.

Why it matters

Video volume decides who wins attention. PixVerse gives you volume without the cost spiral. It is perfect for quick A B tests, scroll stoppers, UGC flavored content, and proving a hook before committing to a real production. When a client hesitates on video budgets, this is the on ramp. You show results first, then scale the spend.

Use it with intention

The output is fast and usable, even if not studio grade. You will see the occasional artifact or weird transition. And because it runs on a credit model, careless testing can get expensive.

Also remember that relying on trending effects can flatten your brand into everyone else’s feed, so push the tool instead of settling for templates.

PixVerse works best as a rapid prototype engine. Use it to explore angles, validate ideas, and build early traction. Then hand the winners to your creative team for polish. That keeps your brand sharp and your testing loop fast.

I like Google VEO 3 better, but it also comes at a much higher price point!

The future of e-commerce requires scale and speed. PPC Ninja helps brands dominate the AI transition. We leverage AI to build stunning, high-converting images and video, efficiently scaling your content production across all channels (Amazon Ads, Social Media, Posts). Stop fighting for relevance. Reach out to support@ppcninja.com to explore how we can immediately upgrade your content and future-proof your listings.

FFMPEG Use cases

I recorded the entire Amazon unBoxed 2025 keynote address (1.5 hours) on my iPhone audio recorder. Later, I asked ChatGPT for an ffmpeg command that would shrink the 50MB+ audio file to the smallest file size possible. A few seconds later I had a tiny 500KB version saved. Simple, direct, efficient. That is ffmpeg in a nutshell.

I then uploaded that tiny file into Google Gemini and had it transcribed and turned into a LinkedIn post with my key conference updates. This is what I call efficiency in the AI era.

So, what is FFmpeg?

A command line tool that reshapes audio and video with surgical precision. Trim it, compress it, convert it, extract it. If your media workflow feels clunky, FFmpeg often has a cleaner way through it. Most web based audio and video editors rely on it behind the scenes. You just never see it working.

Five ideas for using FFmpeg

1) Audio compression
Takes a heavyweight recording and turns it into something light enough for quick sharing without trashing the quality.

ffmpeg -i input.m4a -b:a 64k output.m4a

2) Format switching
When an editor complains about a file, FFmpeg just converts it and moves on.

ffmpeg -i input.mp4 output.mp3

3) GIF creation
Cuts a small moment from video and loops it into a tight, shareable snippet.

ffmpeg -i input.mp4 -ss 00:00:02 -t 3 -vf "fps=12,scale=480:-1" output.gif

4) Video to audio extraction
Saves only the voice or soundtrack so you carry the message without the full video weight.

ffmpeg -i input.mp4 output.mp3

5) Silence trimming
Removes the dead space at the start and end of audio so it feels intentional from the first second.

ffmpeg -i input.wav -af silenceremove=start_periods=1:start_threshold=-50dB:stop_periods=1:stop_threshold=-50dB output.wav

RUFUS got some prominent real estate

Perhaps you noticed it already but RUFUS just grabbed more real estate on Amazon and it is acting a lot more like a true agent than a simple search feature. If you have used tools like Atlas or Comet, you know the pattern. The assistant sits on the side, thinking while you browse. Amazon just adopted that same layout.

Now RUFUS lives in the sidebar and you can watch it work through real steps. It checks your past searches, looks at your cart, scans your lists, and pulls products together. This is not filtering. This is an assistant taking actions inside the shopping flow.

What makes this shift interesting is that none of the outside players can match it. Not ChatGPT Shopping. Not Perplexity. Not Copilot. Not Google. They are all guessing from the outside. RUFUS sits inside the actual session with access to cart signals, account behavior, and the full catalog.

Amazon released their AI Creative Studio at unBoxed 2025

Amazon launched the Gen AI Creative Studio last year, and now they are introducing their creative agent. It is still early days, but you can see where they want this to go.

In the Unboxed keynote, they walked through how a brand could build an STV ad using only AI. The output showed what you’d expect from a first generation system. It had the core pieces in place and you could see the direction, even if the polish is yet to come.

Amazon framed these as “better than nothing”, which I took less as a verdict and more as a signal of intent. This is a starting point. A way to move faster on drafts, concept tests, and early explorations before investing in fully produced assets.

And all of this is happening while audiences are paying closer attention to how AI shows up in creative work. The recent criticism around Coca Cola’s holiday campaign is a good example. People care, and brands are learning how to navigate that.

The upside is still real. Faster cycles, cheaper experimentation, more creative volume. The trick is using AI where it helps and keeping humans where the stakes are high. These tools will improve quickly. The brands that stay thoughtful through the transition will benefit the most.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

Vibe Coding Your Own Chrome Plugins

Ever felt the need to scrape data off Amazon product detail pages?
Maybe you have a VA copy pasting ASINs, delivery dates, review counts, star ratings, or price history. That works. It is also slow and error prone.

Here is the faster path. Vibe coding your own Chrome plugin.

With Claude, you describe what you want, show it where the data lives, and iterate until it behaves. No heavy engineering. Just tight feedback loops.

Chrome Plugins Shelf

Real use cases for advertisers

If you manage dozens, hundreds or thousands of ASINs like we do at PPC Ninja, you already know how much of PPC optimization depends on the context of a product listing. Whether it is in stock, Prime eligible, or backed by fast delivery etc.

You cannot run profitable PPC campaigns for Halloween products that say Arrives after Halloween. Or bid aggressively when your coupon disappeared overnight.

Now imagine having a tiny browser tool that can check all of that for you automatically. That is what a Chrome plugin can do. Instead of opening listings one by one, your plugin gathers those details in the background and gives you the data in seconds.

That kind of visibility helps you make quick, confident bid or budget decisions without waiting on reports or manual checks.

Bookmarklet vs Chrome plugin

A bookmarklet is JavaScript saved inside a bookmark. Click it and it runs on the page you are viewing. It is great for one offs and quick utilities. Long time readers will remember the RUFUS Q&A extractors we built this way.

Well, a Chrome plugin is better for repeatable work at scale. You can

• Feed it 100s of ASINs and let it open pages in the background while you do other work

• Pull consistent fields every time

• Output to clipboard, CSV, or a popup UI

Why not just go Agentic?

I love agentic browsers. They can explore, reason, and take actions that feel almost human. But sometimes you just need consistency. Agentic tools are still evolving, and as of November 2025 they can be a bit unpredictable or slow, often needing a human nudge or babysitting to stay on track.

A Chrome plugin on the other hand gives you the opposite experience. Controlled, fast, and repeatable. You decide the selectors, timing, and outputs, and get the same result every time.

A concrete example

I recently vibe coded a plugin that checks if an item will arrive before or after Halloween. Because if a seasonal product lands on November 1, that is money on fire.

The plugin visits each ASIN, reads the delivery promise, and returns Before Halloween or After Halloween with the date. It runs across a list while I do other work.

Step 1

Step 2

Wait a few seconds or minutes (depending upon how many ASINs you want checked) and see it pop up the results!

Step 3

What should I scrape?

Common Amazon detail page fields worth automating

• Delivery promise or arrival window

• Price and coupons

• Review count and star rating

• Variation attributes like size or color

• Seller or ships from

• Buy Box owner

• Availability and back order notes

Pick only what you need. Fewer moving parts mean fewer failures.

Output choices

• Copy to clipboard for quick pastes into Sheets

• Save as CSV to import in bulk

• Show a popup table for quick triage

Practical notes

As you are vibe coding, you might see mysterious words like manifest.json, popup.js, or content.js pop up. Do not worry. That is just Chrome’s way of labeling files. You do not need to know what they mean. Your AI assistant writes them for you.

A few quick tips

• Keep permissions minimal

• Add small delays between page loads

• Be respectful of Amazon’s TOS and rate limits

• Start private. If you ever publish it, add icons, a short description, and a basic privacy note

The 2 Minute Vibe Code Checklist

1. Decide what you want to scrape
   Example ASIN, delivery date, or star rating

2. Find each field in the browser
   Open a product page, right click the data, choose Inspect
   Copy the relevant HTML snippet or selector and give it to Claude

3. Ask your assistant to code it
   Example prompt

4. Download and load it into Chrome
   Claude will give you a ZIP. Unzip it.
   Go to chrome://extensions, toggle Developer mode, click Load unpacked, select the folder.
   Pin the extension. Test it on 3 ASINs, then 30, then 100.

Done.

Have you tried vibe coding like this? How has your experience been?

AI Music for your Brand with Suno

Suno is one of the most jaw dropping AI music tools right now. I typed a simple prompt “Create a jingle about my face cream product” and it generated a full track with lyrics, and sounds like it came off a professional TV commercial. The voice, the mix, the production all feel SO real. Have a listen…

(Music cover was designed by me in Canva)

Suno allows you enough credits to get your hands dirty on a free plan. It generates 4 versions for each prompt, 2 free and 2 paid (if you choose to upgrade). The free versions are complete with Lyrics that you can edit and remix!

Suno is one of the serious AI music generators out there and there are rumors that OpenAI is coming out with their own version soon.

Suno allows you to Remix and AI music from it’s artists (don’t worry, it’s all legal and only music from artists who have opted in will be available for remix). I am glad that Suno has taken some steps to protect artist copyrights.

When downloading your free AI music, Suno will prompt you with this message that ensures that you are not using the music for commercial purposes. You will have to either Upgrade for commercial licenses or proceed to Download for your internal or non-commercial purposes, like I did.

This is a HUGE creative leap for music. But it is also a preview of what is coming to e commerce when it comes to ownership.

The music industry is already dealing with these forces that will soon shape how brands operate online. Infinite content and identity confusion. Recently this AI cover of Rihanna’s Umbrella sound track went viral. It is unclear which software tool was used to create it. I am pretty sure it was not Suno, because Suno does not allow you to copy well know artists’ music. Do have a listen though, it is insane…

The Marketplace Flood

Spotify is already fighting AI slop with spam filters. Marketplaces will face the same problem as low quality AI listings flood search results. The risk is not just compliance but visibility.

The lesson from music is simple. Build your brand as if rules are coming soon, because they are.

The future of e-commerce requires scale and speed. PPC Ninja helps brands dominate the AI transition. We leverage AI to build stunning, high-converting images and video, efficiently scaling your content production across all channels (Amazon Ads, Social Media, Posts). Stop fighting for relevance. Reach out to support@ppcninja.com to explore how we can immediately upgrade your content and future-proof your listings.

Vibe Coding

I have been doing a lot of vibe coding with Claude and Gemini lately and wanted to introduce a term you will start seeing more often. CLI, short for Command Line Interface.

If you use a Mac, the app is called Terminal. On Windows it is Command Prompt or PowerShell. If you have ever seen a black screen filled with white text and a blinking cursor, that is it. The command line is a text based way to talk directly to your computer. No buttons, no windows, just clear instructions.

It sounds technical, but it is surprisingly useful once you try a few things.

With a few simple commands you can

• move or rename hundreds of files in seconds

• bulk convert CSVs or resize images

• clean up old folders that are taking space

• run small scripts to automate daily routines

• connect to cloud servers or data pipelines without opening a browser

That is why developers love it. It is fast, predictable, and works anywhere.

For those of us experimenting with AI tools, the CLI is where automation quietly becomes real. When you ask Claude to “write a script that downloads 100 images,” it is essentially writing CLI commands on your behalf.

You do not need to master it, but knowing what it is unlocks a lot of small superpowers. Pair it with your AI assistant and you start blending human intent with machine speed.

It is not flashy, but it is where the real magic begins. More on this in future posts…

PPC Masterclass

If you’ve ever felt like Amazon PPC is draining your profits instead of growing them, you’re not alone. I’m excited to invite you to the Expert PPC Masterclass: Master Amazon Ads, happening Thursday, November 6th at 1 PM ET hosted by My Amazon Guy.

Sign up here for free.
When: Nov 6, 2025, 1pm EST

Amazon Unboxed 2025

The PPC Ninja Team is sponsoring Unboxed! If you are going to be there, hit reply and RSVP for a dinner with our team.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

Forget custom GPTs. Forget copy-pasting your brand guidelines into every single chat or project.

Claude just gave its Pro users a secret weapon called Skills, and it fundamentally changes the game. This isn't just a new feature; it's the end of repetitive prompting and the rise of specialized, autonomous AI.

Find it under Capabilities → Skills, and turn it on. Now read on for the fun part…

What Exactly Is a Claude Skill?

A Skill is a permanent, self-activating module of specialized expertise that lives in your Claude account.

Once uploaded, it's always ready to jump in and apply its specific knowledge to any conversation, without you having to remind it or paste a single line of instructions! For example, the Tax consultant, the Marketing genius, the Compliance expert.

🎯 The Real-World Impact

Here’s the difference between temporary instructions and permanent expertise:

• Before Skills: You open a chat, paste the 40-page EU Compliance Checklist, paste your marketing copy, and ask: "Find the risks."

• With Skills: You open any chat and simply ask: "Review this new landing page for EU compliance risks."

The AI automatically recognizes the need, activates your pre-loaded GDPR/CCPA Compliance Skill, and instantly flags the issues against all your proprietary legal rules—without you entering a single instruction.

🔪 The Kill Shot: Why Skills is Better than Your Current Setup

If you're already a power user, you're probably asking: "How is this different from a Project with Instructions?"

It's about Scope and Automation.

The bottom line: Skills eliminate friction. They allow you to deploy a whole team of specialists—Compliance, SEO, Customer Service—all running in the background, ready to jump in without you asking.

🔓 Unlock the Power: Setup in 60 Seconds

Skills are not the default. You have to flip the switch:

1. Go to Settings →Capabilities.

2. Beneath the Experimental section, find Skills under Turn on Code execution and file creation. (It's required for the magic).

3. The Skills (Preview) option activates. Use the Upload Skill button to deploy your expertise.

⚙️ The Blueprint: Building Your E-commerce Specialist

A Skill is just a ZIP file containing a Skill.md document. The most critical part isn't the instructions; it's the metadata. This defines the AI's autonomous trigger.

Once the trigger is set, you load your detailed rules into the Markdown body: your specific legal statutes, internal checklists, and risk profiles. Upload the ZIP, and the expertise is live!

The best thing is, if you don’t know how to build a Skill file, just ask Claude to build some for you with some basic instructions!

The AI Video Wars: China's "Free" Waver 1.0 Model Puts US Subscriptions on Notice

The battle for the future of AI video is heating up, and ByteDance, the company behind TikTok, just dropped a bomb. They quietly released Waver 1.0, a powerful new AI video generator that is surprisingly free and open to the public.

This move immediately elevates the tech rivalry between the US and China from a competition over quality to a clash of core business strategies. Waver's zero-cost entry contrasts sharply with the pay-to-play models of major Western competitors:

Waver's free launch is a declaration that China is committed to the "reach wins" philosophy, putting massive pressure on the existing Western paywall models.

Waver 1.0: Why the Free Model is a Technical Beast

Launched in August 2025, Waver 1.0 is more than just a budget option; it's a technical powerhouse. The model is already challenging top-tier competitors, thanks to its focus on professional-grade filmmaking features:

• Cinematic Quality: It generates high-resolution 1080p video up to 10 seconds long, double the duration of many existing models.

• Narrative Control: Users can now execute multi-scene workflows within a single prompt, directing the AI to cut between different camera angles, like switching from an establishing shot to a close-up.

• Advanced Consistency: The model is highly praised for maintaining object shapes, realistic physics (think water splashes), and natural human anatomy throughout complex motion.

You can currently access Waver 1.0 on a variety of AI Video platforms like ArtAny.

The Competition Responds: Google's Veo 3.1 Upgrade

The market pressure from Waver was palpable. Google responded in October 2025 by launching Veo 3.1, an upgrade focused on enhancing creative control to maintain its paid user base:

• Granular Editing: Veo 3.1 added features like Object-Level Editing (inserting or removing specific items in a clip) and advanced frame transitions.

• Audio Leadership: Veo continues to be the leader in generating native, synchronized dialogue and sound effects alongside the video.

• Ingredients to Video: For e-Commerce brands, upload your product, a face and a scene and boom.

However, as ByteDance gives away high-quality tools for free, the Western strategy of "Monetize First" risks being irrelevant. Once creators establish their workflow with a powerful, zero-cost tool like Waver, it won't matter if the competitors are slightly better; the market shift will have already occurred.

The Real Fight: Reach vs. Revenue

This move highlights a fundamental strategic divergence between the U.S. and China for control of the future of AI.

The U.S. focuses on locking down high-value users early. China focuses on getting the tool into the hands of millions of creators first.

Why Reach Wins the War

When a high-quality, free tool enters the market, the shift is inevitable.

Once creators move their workflow to Waver, the network effect solidifies Bytedance’s control. It won't matter if competitors fine-tune a marginally better model next quarter. The user base—the real currency in AI—has already been claimed.

The market has shifted. Companies betting on paywalls and credit systems just got blindsided by a competitor who prioritizes global dominance over immediate revenue. The pressure on the subscription model is now immense.

The future of e-commerce requires scale and speed. PPC Ninja helps brands dominate the AI transition. We leverage AI to build stunning, high-converting images and video, efficiently scaling your content production across all channels (Amazon Ads, Social Media, Posts). Stop fighting for relevance. Reach out to support@ppcninja.com to explore how we can immediately upgrade your content and future-proof your listings.

We're all chasing AI adoption, but does that traffic actually pay? Early data from a massive new study offers a necessary reality check (and a glimpse into the future).

Juo Kaziukėnas flags key insights from research by Maximilian Kaiser and Christian Schulze, who analyzed 12 months of first-party data across 973 e-commerce sites and over $20 billion in revenue. They investigated how organic Large Language Model traffic (oLLM) performs compared to traditional channels (like paid search, affiliate, and direct).

The Data: AI Traffic Is Small and Soft

The study confirmed what many suspect: traffic referred from ChatGPT currently converts far worse than every traditional channel. For instance, the conversion likelihood from affiliate traffic is 86% higher than from ChatGPT.

Here are the critical takeaways from the study:

• It's a One-Horse Race: ChatGPT accounts for over 90% of all observed e-commerce sessions from LLM platforms. (You can safely ignore the others for now.)

• The Volume Problem: oLLM traffic is still tiny: about 200 times smaller than Google's organic search, accounting for less than 0.2% of all traffic in the dataset.

• The Quality Gap: People converting from ChatGPT are currently doing so at a lower rate than from nearly all other channels.

The Turnaround: Why This Is Only Temporary

The most important finding? The conversion rate gap is narrowing fast.

This data suggests that as users gain experience and trust with LLMs, they are becoming more likely to follow through on recommendations. This trend will be radically accelerated by two factors:

1. Trust Improves Behavior: As people learn to rely on AI for shopping research, their behavioral usage is shaping up to match traditional channels.

2. Instant Checkout: The introduction of native checkout features within platforms like ChatGPT (replacing the multi-step process with a single "Buy" button) is poised to provide a massive, immediate boost to conversion rates.

The conclusion is stark: Today’s LLM traffic is a trickle of low-converting users, but that is merely the testing phase. The platform usage is increasing, the user experience is improving, and the conversion rates are catching up. None of this will look the same in 12 months.

Read Juo’s article on Linkedin

I am representing Amazon Business at the Small Business Summit in Toronto this month. If you need a FREE ticket to attend in person, hit reply and I will send it across.

Yay, I am going live with Norm Farrar and Amazon Ads on Wednesday, October 22nd! Tune into our conversation on YouTube

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

A whole week of people begging for a Sora invite. I’m guilty of it, too. The hype was real, a clever use of scarcity by OpenAI.

And what did we find when the gates finally opened? A chaotic, scrollable feed of AI-generated vertigo. Hyper-realistic animals doing impossible things, uncanny valley characters in grainy footage, and endless remixes of the prompt-of-the-day.

It’s a platform optimized for novelty, where the commoditization of creativity isn't just acceptable, it's the main form of entertainment.

This is where most people will form their opinion of Sora. They'll see it as a bizarre toy for creating memes, a firehose of digital curiosities fenced in by a moving watermark. And they will move on to the next thing.

But that perception creates the single biggest strategic advantage for anyone paying attention.

Because behind this public spectacle lies a completely different tool. Last week, I ignored the craziness of the feed and gave the engine a focused, commercial prompt: "Create an ad for a loreal professional, Absolute Repair hair shampoo, show before and after on a beautiful model with gorgeous long hair. Product has these qualities: L'Oréal Professionnel Absolut Repair Shampoo, Hydrating Shampoo, For Dry and Damaged Hair, Adds Shine and Repairs Damage, with Omega-9 and Proteins, Professional"

And I intentionally chose landscape, not portrait.

The result was a quiet revelation. It produced near-perfect advertising imagery, almost indistinguishable from a million dollar shoot, embodying the exact qualities of an expensive brief. It was a focused, powerful, and strategically valuable tool.

This is the Sora paradox: its identity as a social platform is a massive distraction from its power as a professional one. The mistake is confusing the feed for the engine.

• The Feed is an Echo Chamber: The public platform is designed for TikTok style engagement and entertainment. It rewards speed, imitation, and remixes, creating a "Sora-core" aesthetic. Chasing trends in this echo chamber is a race to the bottom for any serious brand.

• The Engine is a Strategy Lab: Used privately, the underlying engine is a powerful tool for de-risking creative bets. It’s for testing a dozen campaign moods, visualizing a product in an ideal setting, and closing the "imagination gap" between a brief and a tangible asset.

The cute little dynamic watermark plays a crucial role here. It’s a constant reminder that the content generated is native to this new ecosystem. It keeps the viral memes contained while forcing serious users to treat the output not as a final product, but as what it truly is: a strategic blueprint of what is to come.

The traditional photo studio isn't obsolete, but its purpose is being redefined. It’s no longer just for capturing a baseline reality; it’s for executing a highly-refined vision that was first perfected in a private, AI-powered lab.

The real opportunity with Sora isn’t to create the next viral clip that gets lost in the feed. It’s to use its engine to develop a vision so clear and powerful that it wins everywhere else. Ignore the noise. Focus on the signal.

For you geeks out there, OpenAI announced an API to access Sora 2 at the Dev summit today.

Have you tried Sora for your own brand yet?

I've been obsessed with Gemini lately. The platform is a beast, and it has unlocked the freedom to create micro-tools that solve real workflow problems for creative teams.

So, I built one to tackle a universal pain point: cleanly separating a subject from their background.

Meet the Foreground/Background Separator.

You give it any image, and it instantly splits it into two perfect layers: the isolated subject on a transparent background, and the isolated background with the subject intelligently removed.

Why do I need ANOTHER background separator?
Because it’s FREE. Read on..

• Instant Cutouts: Create clean assets for thumbnails, presentations, and social graphics in seconds.

• Generate Clean Plates: Never worry about forgetting to shoot an empty background again. The tool creates a perfect "clean plate" for you.

• Effortless Compositing: Easily place text, logos, or graphics behind your subject for that professional, layered look.

• No Green Screen Needed: Swap out a messy office background for something clean and simple, on the fly.

This is the kind of practical magic that eases workflows and unlocks new creative avenues. It's one of the 20+ tools I've built with Gemini last week, and it won't be the last. You know me ;)

PPC Ninja helps brands future proof their listings for AI, helping you build stunning images and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

If you missed this Monthly Crusher with Ritu Java and Kevin King in BDSC, you can now watch the replay!

(Note: You will need to be a member of the circle to access)

But this is where I pulled back the curtain on my eco-system for building AI tools in-house. If I can do it, you can do it too.

For months, we've talked about "conversational e-commerce" as a future concept, but a post from Andrew Bell shows it’s happening right now within Rufus.

Amazon's AI assistant is no longer just an Amazon-first tool. It's now acting as a "Buy for Me" agent, occasionally including external products and direct "Shop" buttons that lead off-platform.

Let's be clear: this isn't Amazon testing a small feature. This is a calculated power play.

It seems counterintuitive. Why would the ultimate walled garden build a gate and point customers to the exit? Because Amazon isn't trying to win the transaction anymore. They're orchestrating a movement to own the entire starting point of commerce.

This is a strategy to become the indispensable AI layer for all of retail:

• Own the Interface: By becoming the default starting point for every shopping journey, even for things they don't sell, they make the Amazon app the universal gateway to commerce.

• Master the Data: This position gives them priceless data on consumer intent, pricing, and competitors across the web. They learn what you want, even if you ultimately buy it elsewhere.

• Manufacture Trust: By occasionally linking out, Rufus appears objective. It builds trust not in Amazon the store, but in Rufus the agent. That loyalty is far more valuable in the long run.

• Build the Moat: Ultimately, this is a defensive strike to ensure a Google or Apple agent doesn't become the AI layer on top of their business. Amazon is building its own moat for the age of AI.

They are willingly sacrificing the battle for a single sale to win the war for the entire shopping interface. Thoughts?

Scale Summit - AdCelerate Every Stage

Oct 6, 2025, 10:00 AM - Oct 8, 2025, 12:00 PM PST
Sign up: https://www.linkedin.com/events/scalesummit-adcelerateeverystag7373360802478727168/

I am representing Amazon Business at the Small Business Summit in Toronto this month. If you need a FREE ticket to attend in person, hit reply and I will send it across.

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

Most people throw words like “cinematic” or “epic” at an AI video model and hope for the best. The results are usually… generic. A floaty, sterile clip that feels like stock footage from another dimension.

What the pros know is that there’s an actual vocabulary directors and cinematographers have used for a century. When you speak that language, the AI model suddenly understands the assignment. It has been trained on millions of clips tagged with these exact terms.

By using the shorthand of filmmaking, you aren’t just asking for a vibe; you are providing a blueprint. You're moving from a vague wish to a specific command.

Think of this as your secret language for directing an algorithm like a seasoned Director of Photography (DP).

The Cinematographer's AI Cheat Sheet

Here is a grid designed for action. Find the effect you need, grab the term, and build it into your prompt.

Part 1: The Shots (Framing Your Scene)

Start with the fundamental building block: the shot type. This determines the scale and perspective of your scene.

Drone Shot

Part 2: The Angles (Creating a Feeling)

The angle of the camera dramatically alters the power dynamic and mood of your subject.

Low Angle Shot

Part 3: The Movement (Guiding the Eye)

Static shots are fine, but movement brings your scene to life. Use these terms to guide the camera's path.

360 Orbit

Part 4: The Lens & Lighting (Painting with Light)

How you capture light and focus is just as important as the shot itself. These terms control the visual texture.

AI Lens Flare

How to Stack These Terms: From Prompt to Storyboard

The real power comes from combining these terms. You’re no longer asking for one thing; you're describing a sequence.

For exmaple: “Handheld POV chase through a neon-lit alley, whip pan into a drone shot overhead.”

This isn’t a prompt. It’s a storyboard. It gives the AI commands for style (handheld), perspective (POV), setting (neon-lit alley), transition (whip pan), and new perspective (drone shot).

Here are a few more examples:

• For a dramatic reveal: “Low angle shot of a hero on a cliff edge, slow push-in on their face, rack focus to the storm clouds gathering behind them, low-key lighting.”

• For a character introduction: “Tracking shot following a detective through a smoky bar, 360 orbit as they spot a clue on a table, extreme close-up on their eyes widening.”

• For a high-energy sequence: “Montage of a race car driver preparing: close-up on hands gripping a steering wheel, FPV shot from inside the helmet, whip pan to a wide shot of the track, golden hour lighting.”

The difference is night and day. AI models respond to precision. This is your new shorthand for getting exactly what you envision.

Your turn. Grab a few terms from the grid and try building a sequence. Drop your best "storyboard prompt" in the comments. Let's see what you can direct.

PPC Ninja is helping brands future proof their listings for AI, helping you build stunning images and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

You've obviously seen the headlines about ChatGPT enabling in-chat purchases. It's easy to dismiss this as another tech novelty, but that would be a strategic mistake. This isn't just a new feature; it's the blueprint for a new, powerful sales channel that operates on fundamentally different rules.

Here’s a breakdown of the implications for your business and what you should be thinking about right now.

The Strategic Shift: From Destination to Conversation

For years, the goal has been to drive traffic to your website, your digital storefront. You controlled the experience, the branding, and the checkout process.

That model is now being challenged.

With what OpenAI is calling Agentic Commerce, the AI acts as a trusted personal shopper. The customer's journey doesn't start on Google or your homepage; it starts with a conversation. They state their intent ("I need a durable, waterproof backpack for hiking"), and the AI does the work of finding the product and facilitating the purchase. Your storefront becomes a data source for the agent, not the destination for the customer.

This is a critical distinction. You are no longer just selling to a person; you are optimizing your business to be chosen by an intelligent intermediary.

Immediate Implications for Your Business

1. "Agent Optimization" is the New SEO

While your current SEO strategy is built around pleasing Google's algorithm, you now need a strategy for AI agents. These agents prioritize structured, unambiguous data. They need to know product specs, inventory levels, shipping policies, and pricing with absolute clarity. Vague marketing copy is less valuable than clean, machine-readable product feeds. The businesses that make their products the easiest for an AI to understand and transact with will win.

2. Technical Integration Has Become a Competitive Moat

The Agentic Commerce Protocol (ACP) is the open-source standard behind this. The OpenAI announcement explicitly states that when multiple merchants sell the same product, the one with Instant Checkout enabled is prioritized. This means early and seamless technical adoption is no longer just an operational concern, it's a direct competitive advantage. Waiting to integrate could mean losing the "buy box" to a more agile competitor, even if your price is the same.

3. The Definition of "Customer Relationship" is Changing

When a customer buys through ChatGPT, who owns the relationship? You get the order and the shipping details, but the AI owns the discovery and checkout experience. This makes direct brand-building more challenging and more critical. If customers are interacting with a generic AI, your brand's reputation and equity are what will drive them to ask for your products by name. A strong brand becomes one of the few ways to ensure the AI selects you.

Your Action Plan: Three Things to Do Now

1. Conduct a Data Audit. Task your team with a full audit of your product information management (PIM) system. Is your data structured, accurate, and API-accessible? If an AI asked for your product catalog, would it get a clean, comprehensive answer? This is now a mission-critical priority.

2. Evaluate the Agentic Commerce Protocol (ACP). Have your engineering lead investigate the ACP immediately. As outlined by OpenAI and Stripe, implementation is straightforward for businesses already on Stripe. If not, understand the path to integration. The goal should be to get this on your technical roadmap within the next quarter. A first-mover advantage here is significant.

3. Invest in Brand, Not Just Performance Marketing. Reallocate a portion of your marketing budget toward building brand equity. In a world where AI agents can find the cheapest or most convenient option, a brand that people trust and ask for by name is your ultimate defense. Your goal is for the customer to tell the AI, "Find me a backpack from your brand," not just "Find me a backpack."

And if you'd rather focus on product than protocol, leveraging a platform like Shopify is your simplest path forward. They have integrated the technology directly, meaning for most merchants, enabling this new sales channel is a simple activation within the Shopify admin, not a complex engineering project.

This is the starting gun for the next era of e-commerce. It's a shift from visual browsing to conversational command. By treating it with strategic importance now, you can position your business to thrive in it.

If you missed this killer episode of the Go with the Flow Podcast, you can now watch the replay!

Go With The Flow: Agentic Browsers, Claude Code, and the Future of E-commerce Automation

In this episode of "Go With The Flow," I rejoin co-host Danny McMillan to dive deep into one of the most transformative shifts in e-commerce: agentic technology. We explore how new tools like Google's agentic Chrome browser and platforms like Claude Code are moving beyond simple prompts to actively execute complex tasks on our behalf. From automating tedious reports in Seller Central to building entire workflows with conversational commands, we break down how you can start leveraging these powerful systems—even without a technical background.

TLDR;

Stop clicking. Start commanding.

This episode is about ditching the "busy work" that drains your energy. My rule? If you do it 3 times, an AI should do it forever. I show how I made an agentic browser do my tedious report-pulling inside Seller Central.

Then, Danny demystifies the "tech stuff" by revealing how he achieves weeks of work in a single day using Claude Code. The secret isn't being a programmer, it's about being fearless and treating the AI like a new hire. This is your guide to getting the AI to work for you, not the other way around.

Have a listen!

Andrew Bell recently spotted a critical new development that pushes the boundaries of conversational commerce: The "Add to Cart" function in RUFUS. This is the most significant step yet toward true AI agency in shopping on Amazon! Read the full piece here:

PPC Ninja is helping brands future proof their listings for AI, helping you build stunning images and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

I am BEYOND excited and honored to be a keynote speaker at the E-Commerce Fuel Summit this week. I have 2 talks:
1) Turning Browsers into Buyers with AI-Driven Creatives
2) Beyond the Basics: A Deep Dive into AI-Powered Creative Strategies

Scale Summit - AdCelerate Every Stage

Oct 6, 2025, 10:00 AM - Oct 8, 2025, 12:00 PM PST
Sign up: https://www.linkedin.com/events/scalesummit-adcelerateeverystag7373360802478727168/

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

How to Mimic Your Favorite GPT and Make It Run Anywhere

You've found the perfect custom GPT. Maybe it's a brilliant writing editor, a coding assistant that just gets you, or a creative brainstorming partner. The problem? It’s locked inside the OpenAI ecosystem. What if you want that same powerful assistant in Google Gemini, Claude, or another platform?

There's a surprisingly simple method to "mimic" the core functionality of any custom GPT. You don't need to be a developer or a prompt engineering expert. You just need to ask the GPT to explain itself.

This guide will show you how to extract the core instructions from a custom GPT and port them to another AI, giving you a consistent, powerful assistant on any platform you choose.

Why Bother Mimicing a GPT?

Recreating a GPT's instructions elsewhere offers a few key advantages:

• Platform Independence: You're no longer tied to a single service. Use your favorite AI personality wherever you work best.

• Cost-Effectiveness: If your preferred GPT is on a paid plan, you might be able to replicate its core behavior on a free, high-quality model. For example, Google Workspace users don’t need to pay for Gemini Gems.

• Deeper Understanding: This process reveals how custom instructions shape an AI's behavior, making you a more effective prompter.

• Improved Functionality: You can take what you don’t like about the GPT and turn it into something even better.

Claude already natively supports this functionality. Any public Artifact created on Claude can be “Remixed” or “Customized” to generate new and improved functionality.

The 4-Step Guide to Mimicing a GPT

The process boils down to a simple conversation where you ask the GPT to generate its own instructions. This usually works well with simple GPTs that don’t have a ton of steps, branches or that have been trained specially on a lot of documents.

Step 1: Choose Your Target GPT

Select a custom GPT you want to mimic. This works best for GPTs defined by their persona, process, or specific output format. Examples include writing coaches, creative title generators, code reviewers, or role-playing assistants.

Step 2: Prompt the GPT to Explain Itself

Your first goal is to get the GPT to clearly articulate its purpose. Open a new chat with your chosen GPT and use a prompt like this:

This encourages the AI to outline its operational logic, which is the foundation for the next step. Note that not all GPTs are fully open.

Step 3: Ask GPT to draft a Fresh set of Custom Instructions

Now, have the GPT translate its purpose into a set of universal instructions. This is the key to the entire process.

The GPT will now generate a ready-to-use set of instructions that defines its personality, workflow, and guardrails.

Step 4: Transfer and Test

Copy the instructions generated in the previous step. Now, go to your target AI platform (like Google Gemini Gems) and find the custom instructions area.

1. Navigate to your AI's settings. In Gemini, for example, click your profile picture and select "Custom instructions."

2. Paste the instructions you copied into the appropriate field.

3. Start a new chat and test it. Give it the same kind of task you would give the original GPT and see how it performs. You may need to tweak a few phrases for a perfect fit, but the core behavior should be remarkably similar.

A note about Google Gemini Gems. These are just like CustomGPTs. Previously, they were private and could not be shared. But Google just made them shareable.

In recent issues we covered Google's Nano Banana (officially called Gemini 2.5 Flash Image) here and here. And following in the heels of Nano Banana was Seedream 4 (latest, greatest image model from China’s Bytedance).

Now, a new tool called Pippit combines the power of both.

Pippit runs the two image models in tandem, fusing Google's speed and adaptability with Seedream's polished aesthetic. The result is a remarkably versatile and powerful platform for visual creation.

Here’s what it can do:

• Instant Scene Generation: Go from a simple text prompt to a fully-realized world, whether you need a photorealistic moon landing or an anime-style portrait next to your favorite idol.

• One-Click Restyling: Upload a photo and instantly experiment with new hairstyles, makeup, or fashion from around the globe.

• Photo Magic: Sharpen blurry images, colorize old photos, or transform portraits into distinct styles like Pixar or cyberpunk with a single click.

Pippit is a prime example of a growing trend: instead of one model to rule them all, the future may lie in smart combinations that offer the best of both worlds.

To see how they stack up, we ran the exact same prompt through all three platforms. Here are the results.

Product Inspiration

First-up Seedream 4.0

Next up, Nano Banana:

Finally, Pippit:

We’ll let our readers decide which one is the best! Go ahead and zoom into the details and let us know what you decide.

Ready to scale your images with AI? PPC Ninja is helping brands future proof their listings for AI, helping you build stunning, high converting listings and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

If you use Claude regularly, some of its most powerful features are turned off by default. I stumbled upon these settings recently and they're game-changers.

Here are three features you should enable to get more out of the tool.

1. Activate AI-Powered Artifacts

If you build mini-tools with Claude's Artifacts feature, you can now give those tools an AI brain of their own. This allows the interactive apps you create to call on Claude's intelligence directly, rather than just being static outputs.

• Why it matters: Instead of just generating code for a tool, you can build a tool that uses AI to analyze data, generate content, or answer questions within its own interface.

• How to enable it: Go to Settings → Features and toggle on AI-Powered Artifacts.

2. Turn On Long-Term Memory

By default, Claude starts fresh with every conversation. Enabling Memory allows it to recall context and details from your previous chats, leading to more personalized and consistent responses over time.

• Why it matters: Claude will remember your preferences, project details, and writing style, so you don't have to repeat yourself in future conversations.

• How to enable it: Go to Settings → Memory and toggle it on.

3. Connect Your Canva Account

Streamline your creative workflow by integrating Canva directly with Claude. You can brainstorm a complex visual idea, have Claude generate a detailed prompt for it, and then ask Claude to create editable designs directly in Canva.

• Why it matters: This closes the gap between idea and execution, turning your text-based concepts into visual designs in a single, fluid process.

• How to enable it: Look for the option to connect Canva in your Claude settings —>Connectors

  

Honestly the Canva AI connector was surprisingly good. Here is my prompts:

I asked Claude to generate a design with Canva. Claude goes about creating this remote MCP connection to Canva, which you will need to Allow.

Claude creates 4 designs that are editable in Canva.

You can just open them directly in Canva and edit the text overlays easily!

Andrew Bell did it again. He has another great article covering how ChatGPT Is reshaping shopping insights. Check it out:

Marketing Misfits

I am supers stoked to be featured on the Marketing Misfits podcast by Kevin and Norm! We chatted about the impact of AI on e-Commerce and a framework for targeting 5 levels of audience awareness. Here’s the full episode. Have a listen!

Seller Fest Online (FREE)

Excited to be among 20 Live Interactive sessions teaching you actionable techniques to grow your Amazon business.

When: September 23, 2025, 10:30 am PST
Register here for FREE.

Go with the Flow Podcast

Every 4th Tuesday of the month, 8 am PST, join Danny McMillan and me LIVE on the Go With The Flow Podcast

Billion Dollar Seller Club (BDSC)

Are you part of the Billion Dollar Seller Club (BDSC) yet? If not, join it! And then you can make it my 90-minute session on a topic I am super passionate about.

When: September 26, 11 am PST

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

Google's Nano Banana (officially called Gemini 2.5 Flash Image) has taken the AI world by storm, generating over 200 million images in just a few weeks since launch. This isn't just another image generator – it's the tool that's revolutionizing how eCommerce businesses create product visuals, marketing content, and customer experiences. If you missed last week’s Nano Banana deep dive go here.

Now, we know that the world already said things like RIP Nano Banana and moved on to Seedream 4 (latest image model from China’s Bytedance) we are actually betting on Nano Banana as the safe choice and cannot stop obsessing over it. So we decided to give you some more ideas for your e-Commerce images.

1) Product Variation Generation

Create multiple color/style variants without expensive photoshoots.

Prompt: "Take this product image and show it in [forest green/sky blue]. Keep the same lighting, shadows, and background exactly as shown."

2. Seasonal Campaign Assets

Prompt: "Integrate this product photo into a holiday campaign scene – add fall elements like pumpkins"

3. A/B Testing Creative Variants

Prompt: "Integrate this product into a natural outdoor setting and a Modern Lifestyle setting”

BTW, did you notice that Gemini ads a watermark in the bottom right corner? Well, you can actually remove that. All you need to do is create another tool using Claude AI or Gemini code. Just describe it in words and the job will be done (we created one…)

Cool, right?

Ready to scale this? PPC Ninja is helping brands future proof their listings for AI, helping you build stunning, high converting listings and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

📁 Add Chats to Projects - Finally!

For those of you organized souls who like to keep similar chats together in projects, this one's for you. You can now add any chat to an existing project.

Why This Matters:

• Accumulated Training: Keep all your product research chats in one project so the AI learns your brand voice and preferences over time

• Common Instructions: Set project-level guidelines that apply to all chats (like brand guidelines, tone, or specific eCommerce KPIs to focus on)

• Context Continuity: No more re-explaining your business model every time you start a new chat about inventory management or customer analysis

Pro Tip: Create separate projects for different business areas - one for "Product Development," another for "Marketing Campaigns," and a third for "Customer Service Training." Your AI will get smarter about each domain over time.

🌿 Branch Conversations Like a Git Repo

The "Branch in new chat" feature is a game-changer for anyone who's ever been deep in an AI conversation and wanted to explore a different direction without losing their current thread.

How It Works: Simply right-click any message in your conversation and select "Branch in new chat" - you'll get a new conversation that starts from that exact point, while your original chat continues independently.

How to locate branches? They will automatically be marked as a “Branch”

🔇 Turn Off GPT-5's Over-Chattiness

If you're like me and find GPT-5's latest conversation mode annoyingly verbose and overly chatty, you can now turn off "Advanced Voice" and return to straightforward, business-focused conversations.

The Problem: GPT-5's default conversation mode often sounds like it's trying too hard to be your friend, adding unnecessary pleasantries and filler words that waste time in professional contexts.

The Solution: In Settings → Personalization → Advanced → Turn off "Advanced Voice" to get back to:

• Direct Answers: No more "Oh, that's such a great question!" before actually answering

• Professional Tone: Responses that sound like they came from a competent business consultant, not an overeager intern

• Faster Interactions: Less verbal fluff means quicker time to actionable insights

My good friend and podcast co-host Danny McMillan produced a GIGANTIC research document explaining MCP (Model Context Protocol) in an easy to understand manner. This is your chance to dive deem into it! He’s spend a lot of time putting this together and it is free! Here’s the full article.

The Death of "Trust Me, You'll Want This in 8 Months"

Retail guru Joanna Williams just called out the elephant in the room: while brands are still planning like it's 1995, consumers have gone full AI-native.

Here's the brutal math: There's $1.7 trillion worth of stuff sitting in warehouses that nobody wants, while the products people actually crave are nowhere to be found. Why? Because brands are still betting on what you'll want next spring while you're already onto three different micro-trends.

Enter Google's Nano Banana: What looks like a fun AI toy is actually the supply chain killer app. Williams sees it as the bridge that collapses parts of the design-to-market pipeline. Instead of waiting months for focus groups, you can generate multiple product variations and test concepts digitally before committing to manufacturing.

The Reality Check: Look, we're not talking about magic here. You still need factories, you still need fabric, you still need shipping. But Williams' point is that the creative and decision-making parts (the stuff that usually takes months of meetings and revisions) can now happen in days. The AI handles the concept generation, visualization, and initial market testing that used to eat up half your timeline.

What's Actually Happening: Brands like Shein already prove this works. They test thousands of micro-designs, manufacture only what shows early traction, and restock winners in weeks, not seasons. The AI tools just make this approach available to brands that don't have Shein's massive digital infrastructure.

The New Rules: It's not about eliminating manufacturing time. It's about eliminating the months of indecision before you even start manufacturing. AI lets you fail fast and cheap in the digital world before you fail slow and expensive in the real one.

Read full post here. She’s on point.

Seller Fest Online (FREE)

Excited to be among 20 Live Interactive sessions teaching you actionable techniques to grow your Amazon business.

When: September 22-25, 2025
Register here: https://sellerfest.online/

Billion Dollar Seller Club (BDSC)

Are you part of the Billion Dollar Seller Club (BDSC) yet? If not, join it! And then you can make it my 90-minute session on a topic I am super passionate about.

When: September 26, 11 am PST

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu

See this comparison between GPT5 and Nano Banana? Obviously Nano Banana wins this round. No extra fingers, no duplicate mascara brush. More natural looking model.

Even when GPT-5 is given the same prompt, it still struggles with details like hands and props (btw, can you spot the 6th finger?). The question isn't which AI model to choose, but how quickly creative teams can adapt their workflows to now leverage Google's superior algorithm.

🍌 The Nano Banana Prompting Formula That Turns Basic Products Into Lifestyle Gold 💥

Google’s Gemini 2.5 Flash (aka Nano Banana) is FREE right now, but most people butcher it with lazy prompts. In this post I share the formula that cranks out $500 looking lifestyle images in minutes. But first some example…

Action Sports Scene Environmental Context

Product: Camping light
Weak prompt: “camping light on hiking backpack”
Nano Banana Formula: “Professional lifestyle photograph of hiking gear at a mountain campsite during golden hour. Dramatic sunset, layered mountains, campfire glow. Backpack placed naturally with boots and accessories. Cinematic National Geographic style.”

Why It Works: Context + atmosphere + lighting + pro reference.

Here’s another example of packing cubes:

Weak prompt: “packing cubes in hotel room”
Nano Banana Formula: “Create travel-focused convenience shot keeping product's precise geometry and text unchanged: Airport, hotel room, or travel setting background. Product in carry-on bag, airplane tray, or hotel room setup. Lighting matches travel environment (fluorescent, window, or warm lamp). Emphasize portability and convenience for travelers. 2000x2000 px square; include passport, boarding pass, or travel accessories as props. Preserve exact form, all components, and structural elements exactly as source.”

Why It Works: Target use case + audience context + style.

Let’s take a supplements example:

Weak prompt: “supplement bottle on kitchen counter”
Nano Banana Formula: “Morning kitchen with supplements among fresh ingredients. Marble counter, fruits scattered, golden sunlight streaming. Wellness aesthetic, daily routine moment. Food photography style.”

Why It Works: Time + lifestyle signals + pro reference.

The Multi-Layer Nano Banana Formula 🧠

1. Foundation: “Professional lifestyle photo” reference styles (NatGeo, Architectural Digest) pro lighting sharp focus. Refer to my previos post on AI art direction here.

2. Environment: Place + time of day + atmosphere

3. Demographic Signals: Examples: Busy mom, fitness enthusiast, young professional

4. Usage Context: Product solving a real problem in scene or product only. Add any overlay text that you would like and how.

5. USPs: Highlight unique product strengths benefits differentiators competitors lack (BPA free lightweight design portable system). These can be taken from your title/bullet points

6. Negative Prompts: Tell the AI what not to do. For example:
   - No product distortion, modification, or artistic interpretation
   - No text alteration, blurring, or logo changes
   - No shape deformation or proportion changes
   - No overly saturated colors that change product appearance
   - No heavy filters or effects that obscure product details
   - No creative liberties with product design or branding
   - Avoid cluttered compositions that reduce product visibility

7. Composition: Tell the AI how to frame it

   Wide shot → product in full environment
   Medium shot → product in use
   Macro detail → quality close up
   Shallow depth of field → lifestyle blur effect

Advanced Prompting Techniques 🎯

Here are some more we more advanced ones we built with Nano Banana:

The "Problem-Solution" Prompt Structure

Instead of just showing the product, show the transformation:
Formula: "[Setting] showing [problem] being solved by [product] for [demographic]. [Emotional outcome] visible in scene composition."
Example: "Cluttered travel scenario transformed into organized packing system using premium packing cubes. Relief and confidence visible in neat, systematic organization. Business traveler aesthetic."

The "Emotional Anchoring" Technique

Every demographic has core emotional drivers. Example:
- Busy Mom: Efficiency, family time, making life easier
- Fitness Enthusiast: Achievement, adventure, pushing limits
- Young Professional: Success, sophistication, work-life balance
Embed these into your environmental choices and scene composition.

The "Competitive Edge" Prompts

Analyze competitor images and prompt for what they're missing:
- More authentic settings vs sterile backgrounds
- Better lighting vs harsh flash photography
- Emotional context vs product-only focus
- Lifestyle integration vs isolated product shots

If you follow this prompting guide you are likely to get 90-95% of the final look with minor pending edits, and let a graphics designer touch up the final bits.

So how to normal people like you and I come up with the detailed prompts that feed image generation models? We don’t. We ask AI to do that job of expanding prompts from simple to complex.

In fact, that’s where my team is building out interactive apps that do this beautifully and predictably.

Want Google’s official guide on Nano Banana?
https://ai.google.dev/gemini-api/docs/image-generation

Ready to scale this? PPC Ninja is helping brands future proof their listings for AI, helping you build stunning, high converting listings and videos with AI. Reach out to support@ppcninja.com to explore how we can scale your content production across Social media, Amazon ads, Amazon Posts efficiently and affordably.

How to Force Nano Banana to Generate the Right Size

One of the biggest frustrations with Nano Banana is its stubborn refusal to respect your size specifications. You can clearly tell it to generate 2000x2000 images for Amazon listings, but it completely ignores your instructions and sticks to whatever aspect ratio your reference image happens to be.

This creates a real problem when you need specific dimensions for different platforms - square images for Amazon, landscape for Facebook ads, or portrait for Instagram stories. No matter how explicitly you specify the target size in your prompt, Nano Banana seems to have a mind of its own.

The Image Resizer Solution

That's exactly why we built an Image Resizer tool. Instead of fighting with Nano Banana's quirks, you can now:

1. Upload your reference image - whether it's landscape, portrait, or already square

2. Select your target dimensions - 2000x2000 for Amazon, or any custom size you need

3. Choose your resize method - fit with white padding to maintain the original image integrity, or crop appropriately

4. Get the perfect reference image - now properly formatted for Nano Banana to respect

The Smart Workaround

Here's the hack that actually works: First convert your reference image to the aspect ratio you want, THEN run your Nano Banana prompt.

When Nano Banana sees a square reference image, it generates square outputs. When it sees landscape, it generates landscape. The tool forces Nano Banana to honor your intended dimensions by giving it a reference image that's already in the correct format.

This simple preprocessing step eliminates the guesswork and saves you from the endless cycle of regenerating images that come out in the wrong dimensions.

Want our Image Resizer tool for free? Share this newsletter with 5 people using your UNIQUE referral link at the bottom of this page. Once 5 people sign up through your link, you'll get instant access to the tool.

Your referral link is waiting for you at the bottom - start sharing and start resizing!

My good friend Emma Schermer Tamir recently shared some insights that totally confirm what I've been thinking about the future of e-commerce. She's convinced that the search bar, you know, that little box we've all been obsessing over for years, is basically dying. Emma points to how we're witnessing three massive shifts: discovery-first platforms like TikTok where people stumble onto products they never knew they needed, AI shopping assistants that actually help you decide what to buy, and direct checkout that cuts out the middleman entirely. Her advice? Stop obsessing over keyword stuffing and start thinking about whether AI can actually understand what you're selling.

Andrew Bell spotted that Rufus now occupies the entire screen on Amazon's mobile app, not just a sidebar. As he puts it, "A9 is still the engine, Rufus is a new dashboard." His point is brilliant: A9 gets you found, but Rufus decides what gets surfaced in conversational flows. Andrew's recommendation? Layer conversational optimization on TOP of your keyword work because we need to prepare for a fundamental behavior shift where customers interact with AI assistants rather than traditional search. Amazon is actively training shoppers to use Rufus through prompts and voice features, making it effortless. Both perspectives validate what I've been saying about where retail is headed, we're moving from keyword-driven discovery to AI-powered conversations.

I’m excited to share that I’ll be speaking at the Amazon Accelerate Pre-Conference Workshop for partner agencies and brands.

Attending Amazon Accelerate? Join me for dinner (FREE, limited seats). RSVP here.

📅 September 15, 2025, 6:00 PM
📍 Seattle (restaurant details will be mailed to you)

Seller Fest Online (FREE)

Excited to be among 20 Live Interactive sessions teaching you actionable techniques to grow your Amazon business.

When: September 22-25, 2025
Register here: https://sellerfest.online/

We hope you liked this edition of the AI for E-Commerce Newsletter! Hit reply and let us know what you think! Thank you for being a subscriber! Know anyone who might be interested to receive this newsletter? Share it with them and they will thank you for it! 😃 Ritu