Greetings from Team CogniScope 👋

CogniScope partners with game studios and publishers to solve their burning operational problems with custom AI builds. We pull and push information across the stack so teams can ask, ideate, plan, and act efficiently, helping them focus on what matters most: building amazing games passionately as they should be.

The publisher-audience-finding study we promised last week has landed. The bet does not become a new product. It becomes a different sales motion: we sell the consulting work first, and let the product surface from real customer projects. Alongside this, we are adopting a structured Demand Side Sales cadence with a hard target of five customer-discovery meetings every week, starting now. Same teams, same tools, same problems we have been working on. A different door.

All information shared in this newsletter is confidential, please do not share or forward.

📣 The Ask

If you have ever watched a small team pivot from product-led to services-led-with-product-on-top, what was the leading indicator that told you the pivot was working versus quietly masking a deeper problem? Two paragraphs by email beats a 30-minute call. This is the call we made this week, and we want the version of it where it goes wrong.

💼 What's moving

• The consulting-agency positioning is the real outcome of last week's study. When we narrowed our broader competitive scan to the slice that is actually relevant to publisher-side audience-finding, the subcategory itself came back densely populated: attribution vendors, listening platforms, and game-specific data players all converge there. More usefully, the team's own conversations with publishers revealed a different pattern altogether: publishers are not asking for one more tool, they are asking for someone to sit in the seat they cannot afford to hire for and figure the problem out alongside them. So instead of redirecting engineering capital toward a new product in that space, we are reframing how we walk into rooms. The new website assembles around this framing and ships this week.

• Demand Side Sales is not new for us in spirit. What is new is running it as a systematic weekly cadence. We have been demand-aware in our discovery for a while and have done many things right on the push side. What we had not done was dissect the pull properly. Even strong push work is structurally weak without that underneath it, and the sales motion as a whole stays one-legged. That is the gap we are closing now. Every customer conversation feeds the next, anchored by a Pool Framework for discovery and a Use Case Framework for solution scoping. The hard target is five customer-discovery meetings every week, starting this week. Three named segments are already in the pool: a publisher segment, a returning mid-studio engagement, and one prospect re-engaged via the Loom motion we ran last week.

• Our paid outbound platform is being retired this week. Email as a channel was already dead for us in this segment (zero meetings and mostly opt-outs across the quarter), and the platform itself was the last vestigial piece of the old motion. Data exported, subscription cancelled, LinkedIn-only outbound continues at the 100-connections-per-week cap. Named-account work replaces broad campaign volume.

📚 What we are studying

Whether Demand Side Sales as a structured weekly cadence (five real conversations a week run on a consistent framework) measurably changes the quality of signal we get from prospects compared to our prior more ad-hoc discovery. The first check is two to three weeks out: are problems landing through a consistent shape, are objections clustering in interpretable patterns, and is the pool of named segments converging on a wedge or fragmenting. If yes, the consulting-agency motion earns the engineering reinvestment. If not, the framework forces the failure into visibility early, which is the real reason we picked it.

Two weeks ago we said we could not land on the right thing to change in our fundamental ways of working. This week we did, in a way that felt small while it was happening. The answer was not "build something new" but "walk in differently." Same teams. Same tools. Same problems. A different door.

If you'd like to quickly catch up let's schedule a 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Taking a step back to see the big picture again.

Greetings from Team CogniScope 👋

CogniScope partners with game studios and publishers to solve their burning operational problems with custom AI builds. We pull and push information across the stack so teams can ask, ideate, plan, and act efficiently, helping them focus on what matters most: building amazing games passionately as they should be.

This week we found a strong new signal. Six distinct problem bets have surfaced in our work till now, and after a publisher conversation this week, the bet around finding the right audience for games has the most energy of the six. Before we redirect engineering capital toward it, we are committing a few days to the business-side study to understand the publisher market and its competitive landscape. Committing to the loudest signal is not the failure mode we are guarding against. Jumping on it without the business work first is. Bullets are limited, so we are using the signal in new campaign messaging while we do the study, not the build.

All information shared in this newsletter is confidential, please do not share or forward.

📣 The Ask

If you have ever watched a small team miss a window by being too cautious, tell us about that one. Two paragraphs by email beats a 30-minute call. We are spending this week on study, not engineering, and we want the version of the case where that decision goes wrong.

💼 What's moving

• To test whether old conversations could be reignited with the right material, we recorded a 2-minute video showcase of the Root Cause Analysis product mockup with a tight story over it and sent it to old leads who had previously shown interest in CogniScope. Five unique viewers so far, a few replied, and threads silent for weeks reopened. No commitments yet, but old prospects who had gone quiet are talking again.

• Email as an outreach channel is dead in our segment. Zero meetings this quarter, mostly opt-outs. A scan of our own inboxes and some new research efforts told us why: a typical professional in gaming industry (may be applicable to other industries)nowadays receives roughly 30 AI-generated outbound emails a day, so anything from an unknown sender goes unread. LinkedIn is now our only working channel, capped at 100 connections per week, which means every prospect has to be evaluated harder rather than spray and pray. For any of you running outbound campaigns yourselves, are they still working, or are you seeing the same?

• AWS approved $25k in credits. GCP extended their existing credits by a year, with Gemini-only coverage on the extension (no third-party models). Combined with our existing Azure pool, this gives us credit-backed access to flagship models across all three major providers (Claude on AWS, Gemini on GCP, OpenAI on Azure), which keeps optionality open as the model landscape shifts.

• Our website is in final assembly with a new framing. The current site presents CogniScope as a packaged product, but the reality is that we are running parallel campaigns aimed at the burning operational problems of the game industry and its studios, not building a single product right now. The new framing makes the custom-problem-solving model explicit so first-time visitors are not confused, and so anyone reading this (that may be you :)) can refer us to someone they know in the easiest possible way.

📚 What we are studying

The bet is narrow: helping game publishers find the right audience for the games they have already published or plan to publish. The study runs over the next few days. We look at the competitive landscape in publisher-side audience tools, where publishers actually spend, what roles they hire for, what data they integrate, and what they leave on the floor. The output is a go or no-go on whether this audience-finding bet is the natural home for the energy. If go, it earns engineering time. If no-go, it dies on paper this week and we lose a few days instead of weeks.

We are aware that a lot of these investor newsletters have started to sound similar (same structure, same tone), since just about everything on the internet right now is at least AI-assisted. We are trying to change things up a bit so that if you only read the lead and the ask, you already know we are real and we are thinking, even before you know what we are doing.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Strategy sessions are better with food

Greetings from Team CogniScope 👋

CogniScope is the role-based AI Sidekick suite for game studios. It removes the repetitive, operational work, so teams can ask, ideate, plan, and act from a single conversation. The system pulls information and pushes execution across the stack, helping game teams to focus on what matters most: building amazing games with their vision and creativity that players will love.

This week we did four extended strategy sessions, a full forensic analysis of every campaign since the start of this year, and an honest conclusion: the problem we're solving is validated, but not yet burning. The analysis also surfaced a new approach we haven't tested, one that turns our existing data pipeline into the outreach itself.

All information shared in this newsletter is confidential, please do not share or forward.

📣 Asks

• If you know marketing or community leaders at mid-to-large game studios running live-service titles, warm introductions remain our highest-leverage ask as we shift outreach toward this segment.

• We're looking for game industry veterans willing to watch a 90-second Loom walkthrough of a real incident analysis and give honest feedback on whether it would have saved them time. One short video, one honest reaction.

• If you have experience with signal-based outbound sales (monitoring real-time events to trigger timely outreach rather than cold campaigns), a 15-minute conversation would help us validate our new approach.

💼 Latest News

• Completed a full strategic retrospective spanning four extended sessions and every campaign since February. The analysis covered 1,951 contacts across 10 campaigns with 41 responses and 18 interested. The conclusion: the root cause investigation problem is real, validated by prospects at Sony, EA, Rockstar, Zynga, DICE, 2K, and Indigo, but nobody described it as acute enough to buy.

• Had a 53-minute deep dive with a Director of Product Development at DICE who validated every aspect of our value proposition. She stayed 23 minutes past the scheduled time, described exactly the pain we're solving, and introduced Self-Determination Theory as a framework for interpreting player behavior that no current analytics tool provides. But she also self-identified as an influencer, not a decision-maker, and showed professional interest rather than urgency.

• Discovered a new signal detection method using our existing 39-game data pipeline. By correlating community engagement volume with concurrent player counts, we can identify real incidents without any AI processing. When both metrics spike simultaneously, it's a definitive signal that something happened. This turns passive data collection into an active incident radar which we can use for specific outreach.

🛠 Product

KPI Investigation Prototype

• The investigation mockup was reviewed and approved by the whole team. Users who saw it in meetings responded positively to the report format and how it surfaces context across data sources. Next step: record short Loom walkthrough videos using a well-known game's real incident as a case study.

• These Loom videos serve a dual purpose. They become both a product demo and an outreach tool. Instead of emailing someone to ask if they have a problem, we send a 90-second video showing a real analysis of a real incident at their company. The outreach IS the product demonstration.

Incident Detection System

• Discovered that community activity volume alone is a reliable incident signal. In our 39-game dataset, one title showed engagement jumping from a baseline of 300-400 interactions to 5,000 in a single week before returning to normal. That spike, visible in raw numbers without any deep analysis, corresponded to a major update that triggered community backlash.

• When community volume spikes correlate with a concurrent player count change, the signal becomes definitive. We're building a trigger-based alert system with three layers: anomaly triggers, data volume alerts, and variance alarms. This transforms the pipeline from a data warehouse into an early warning system.

🤝 Sales

We spent this week analyzing every outreach conversation we've ever had.

• Campaign forensics revealed a clear channel split. Across all campaigns, 62% of LinkedIn replies expressed genuine interest compared to just 13% of email replies. The last two email campaigns produced zero positive responses. All four booked meetings came through warm LinkedIn connections, not cold email.

• Active campaigns span three tracks. A proactivity campaign (263 contacts, 740 emails, winding down with 85% finished), a marketing content email campaign (290 contacts, 518 emails, 245 still in sequence), and a marketing content LinkedIn campaign (100 contacts, 93 active). Aggregate email reply rate sits at 0.24%, though open tracking is disabled for deliverability, making reply rate the most reliable engagement metric available.

• We figured out that cold emailing is a very bad outreach method due to ease of putting up a campaign together nowadays. Our leads receive at least 20 AI generated outbound emails everyday which leads them to immediately skip any emails coming from strangers. Unfortunately linkedin limits the number of connection request per week so only option left for us to improve our campaign is to evaluate leads much stricter. This requires a lot of detailed work and system since spray and pray do not work anymore with only scaling the outbound.

📊 KPIs

Validation Metrics

Measuring validation strength through:

• Meetings with budget-holding stakeholders

• Escalation depth (does contact bring someone more senior?)

• Speed of problem pull within sprint cycle

• Quality of objection patterns in replies

Sprint Targets

• Meaningful response rate: 5%+ (current email aggregate: 0.24%; LinkedIn data not available via API)

• Discovery meetings booked: 3+ (DICE follow-up pending, one prospect returning from vacation, two warm leads re-engaged this week)

• Escalation: at least 1 contact brings a colleague or superior into a follow-up

• Loom outreach: record and send first signal-based incident video

👷 Challenges

• After 1,951 contacts across 10 campaigns, nobody has shown genuine urgency to buy. 41 people responded, 18 expressed interest, and several validated the problem with detailed descriptions of their investigation workflows. But even the most engaged prospects describe root cause investigation as a manageable annoyance, not an emergency. The team is testing whether this reflects market reality or a ceiling imposed by discovery-style messaging that asks "is this a problem?" rather than demonstrating a solution.

• Email as a channel is producing diminishing returns. The proactivity email campaign's reply rate sits at 0.14%, down from the initial campaign's 1.0%. Meanwhile, 53% of email replies across all campaigns were hostile (opt-outs or requests to stop) compared to 8% on LinkedIn. With LinkedIn capped at roughly 200 connections per week across two accounts, we haven't yet found a replacement for email's volume.

• Cloud infrastructure credits clarified but reduced from expectations. GCP credits were extended to April 2027 but represent approximately $93K in remaining balance, not the $350K+ initial communications suggested. The team is evaluating AWS credits ($25K via Techstars) and open-source model alternatives to reduce cloud dependency.

We spent this week reading 1,951 data points and went back to our previous meetings; The pattern that emerged wasn't about messaging or channels. It was about approach. Every campaign we've run asked some version of "is this a problem for you?" The one time we showed someone their problem instead of asking about it, it reopened a conversation that had been cold for seven weeks. We still don't know if that scales. But it's the only motion we haven't properly tested, and the data pipeline that's been quietly collecting for months might be exactly the tool that makes it possible.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Filiz waving at cold prospects.

Greetings from Team CogniScope 👋

CogniScope is the role-based AI Sidekick suite for game studios. It removes the repetitive, operational work, so teams can ask, ideate, plan, and act from a single conversation. The system pulls information and pushes execution across the stack, helping game teams to focus on what matters most: building amazing games with their vision and creativity that players will love.

Seven weeks of silence on Riot Games. This week we broke it, not with another meeting request, but with a concrete product demo showing we went back to the drawing board. Meanwhile, a new prospect appeared and a fresh campaign targeting marketing leaders launched with 390 contacts.

All information shared in this newsletter is confidential, please do not share or forward.

📣 Asks

• If you know marketing or community leaders at mid-to-large game studios running live-service titles, warm introductions remain our highest-leverage ask as we shift outreach toward this segment.

• We have a design partner prospect exploring KPI anomaly detection for gaming. If you know anyone with hands-on experience running analytics or business intelligence at a game studio, a quick intro could help us validate the prototype faster.

• We're evaluating AWS Bedrock as our primary cloud platform through the Techstars credit program. If you've navigated this process or have experience with Bedrock specifically, a 15-minute conversation would be valuable.

💼 Latest News

• After seven weeks of silence on Riot Games, we re-engaged with a product-first approach. Instead of requesting another call, we recorded a video walkthrough of a real root cause investigation using live LoL data: a champion select crash signal surfaced, and the demo showed how the system deep dives into the cause and recommends action within minutes, not the days it typically takes waiting for another department's report.

• A new discovery prospect emerged through a team member's network, leading to a 45-minute deep dive on anomaly detection for gaming. A follow-up demo is scheduled for this week. This is our first real design partner signal in weeks.

• A new outreach campaign launched targeting marketing decision-makers at large studios, with 390 contacts across email (290) and LinkedIn (100). The messaging focuses on measuring business impact of marketing content, a pain point that kept surfacing in replies from marketing teams in previous campaigns.

• Completed an AWS Bedrock evaluation through the Techstars startup program ($25K credits, 2-year term). With our current cloud credits approaching expiration, this is being evaluated as a migration path.

🛠 Product

Riot Case Prototype

• Built and shipped a mock prototype for root cause investigation. The demo uses a real LoL scenario: a signal surfaces (champion select crashes spiking, players losing LP), and the system traces the root cause and surfaces actionable context in minutes rather than waiting days for a cross-department report.

• The prototype was recorded as a video walkthrough and sent directly to Riot decision-makers, making it forwardable to internal stakeholders who weren't in earlier conversations.

Ongoing

• KPI product proactivity revision started, refining how the system surfaces actionable insights before users ask for them.

🤝 Sales

The biggest development this week wasn't a new contact. It was reviving the Riot Games relationship.

• Re-engaged Riot after seven weeks of silence. The comms side received a video demo showing our new root cause prototype. The engineering side received a status update. Both touchpoints were designed to be forwarded internally to senior stakeholders.

• New campaign (Business Impact of Marketing Content) launched with 390 contacts. Email: 290 contacts, 166 delivered so far, still ramping. LinkedIn: 100 contacts, 97 active in sequence. Too early for meaningful reply data.

• Proactivity Acting Fast campaign continues with 263 contacts, 601 emails delivered, 1 reply, 161 contacts still active in sequence.

• Legacy campaigns winding down. The mid-sized studio email campaign (338 contacts, 9 total replies) and the original proactivity validation (232 contacts, 2 replies) have essentially concluded.

• DAU integration into the data scheduler is underway, enabling us to detect player count anomalies and use them as outreach signals. When a game shows unusual activity, we can reach prospects with relevant, timely context instead of cold messaging.

• Warm lead re-engagement planned for next cycle: reviewing previous discovery meetings (Miniclip, EA, and other contacts from earlier rounds) to identify whose pain points align with current campaigns.

📊 KPIs

Validation Metrics

Measuring validation strength through:

• Meetings with budget-holding stakeholders

• Escalation depth (does contact bring someone more senior?)

• Speed of problem pull within sprint cycle

• Quality of objection patterns in replies

Sprint Targets

• Meaningful response rate: 5%+ (previous campaigns: 5.8% email, 14.4% warm LinkedIn; new campaign still ramping)

• Discovery meetings booked: 5+ (1 confirmed this week, re-engagement tracks pending response)

• Escalation: at least 1 contact brings a colleague or superior into a follow-up

👷 Challenges

• Cloud infrastructure costs require a decision by June. Current credits expire without an extension option at ~$15K/month burn rate. The AWS path ($25K credits, 2-year term) is structurally better but requires migration effort. Delaying the decision compresses the timeline further.

• The Riot re-engagement is high-stakes and binary. If neither track responds within ~10 days, we're back to square one on our strongest lead. The timing is deliberate: one contact returns from medical leave this week, which may create a natural re-engagement window.

For months we've been asking studios to tell us about their problems. This week we flipped it: we showed Riot what a solution looks like using their own data, this used to take a lot of time and effort so we were trying to validate hard before building. The shift from "validate our hypothesis" to "here, look at this" feels small but it may change the entire conversation. People engage differently when there's something on the table to ideate, and since the cost of development is very low right now, we are thinking to do more for each solution idea without waiting for people's response but to show them the solution idea real quick and get feedback early on. One of the biggest learning is the industry is not solution aware but problem aware.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Greetings from Team CogniScope 👋

Too funky for this neighborhood.

CogniScope is the role-based AI Sidekick suite for game studios. It removes the repetitive, operational work, so teams can ask, ideate, plan, and act from a single conversation. The system pulls information and pushes execution across the stack, helping game teams to focus on what matters most: building amazing games with their vision and creativity that players will love.

This week forced some honest conversations. Two campaigns ran to conclusion and the data didn't tell us what we hoped. Mid-sized studios barely responded, and the replies we did get from larger studios were polite but lacked urgency. The clearest signal came from an unexpected corner: marketing and community teams, not product teams, kept describing the problem as real and underserved. We're adjusting.

All information shared in this newsletter is confidential, please do not share or forward.

📣 Asks

• If you know marketing or community leaders at mid-to-large game studios (particularly those running live-service titles), warm introductions would be incredibly valuable as we pivot outreach toward this segment.

• We continue to look for game industry veterans willing to share their operational pain points informally, especially around crisis response timing and content impact measurement.

💼 Latest News

• We completed a full strategic retrospective of all outreach campaigns to date, spanning two months and ~900 contacts. The conclusion: mid-sized studios (50-200 employees) did not resonate with our "KPI shift" framing, but marketing and communications people at large studios showed consistent interest across all channels.

• A new campaign launched mid-week targeting 263 contacts with revised messaging. We realized previous campaigns were essentially asking people to validate a problem for us, which produced polite replies but weak intent. The new approach invites prospects to co-build a solution. The logic: if someone is willing to spend time with a stranger to ideate on a fix, the problem is probably burning. We're testing that assumption now.

• A new discovery meeting was booked with EA after a prospect confirmed that KPI investigation challenges are real at their scale, and a second meeting is scheduled with Rogue Fury in two weeks.

🛠 Product

Data Pipeline

• The ingestion pipeline was restructured, now processing 39 games in under one hour compared to 15 games in three hours previously.

• Every platform's data (Steam, Reddit, Google, etc.) is now tracked independently per game with automatic retry logic and fallback mechanisms. If one platform fails for a specific game, the pipeline logs the failure, skips it, and picks it up automatically on the next scheduled run.

• Database schemas were fragmented and re-indexed so each data source populates its own columns cleanly, replacing the previous monolithic structure. Over 5,000 community posts processed per pipeline run.

🤝 Sales

With two campaigns running to conclusion and data challenging assumptions, the team used this week to question the approach and redesign the next wave.

• Campaign 1 (KPI Anomalies, large studios) produced the strongest signal: product people said "we have teams for this," but marketing and community contacts described the problem as urgent and underserved. Content impact measurement and crisis response speed emerged as recurring themes.

• Campaign 2 (Mid-sized studios, 50-200 employees) reached ~494 contacts across email (338) and LinkedIn (156) with 14 replies total (9 email, 5 LinkedIn). The "KPI shift" terminology was too abstract for smaller studios. Feedback was blunt: "What do you mean by KPI shift?" The campaign confirmed that this specific framing does not transfer to the mid-market segment.

• Key discovery insight: the problem is not detection. Studios already monitor everything with dashboards, social listening tools, and dedicated analytics. The pain point is speed of root cause analysis. Riot described needing a "good enough" analysis in the first 12 hours, because waiting 48 hours for the insights team means the damage has already compounded. Rockstar echoed a similar gap around confidence and clarity in crisis windows.

• Messaging evolution: earlier campaigns validated the problem but replies were often polite rather than urgent. People confirmed "this is real" without it being burning enough to act on. The new campaign shifts the CTA from "tell us about your problem" to "build the solution with us." If someone is willing to invest time with a stranger to co-design a fix, that signals a genuinely burning problem, not just a recognized one.

• Strategic pivot decided: the next campaign will target marketing and community decision-makers at large studios with a "root cause speed" angle, building on this co-creation approach.

• CRM consolidation progressed with a plan to establish a single source of truth across outreach, email, LinkedIn, and meeting data. Over 6,700 contacts loaded, with campaign tagging and deduplication scripts being improved.

• Outreach volume optimization: email targets reduced from 360 to 300 per cycle, with the overflow (~60 contacts) shifted to LinkedIn where follow-up messaging has no daily send limits.

📊 KPIs

Validation Metrics

Measuring validation strength through:

• Meetings with budget-holding stakeholders

• Escalation depth (does contact bring someone more senior?)

• Speed of problem pull within sprint cycle

• Quality of objection patterns in replies

Sprint Targets (3-week cycle)

• Meaningful response rate: 5%+ (previous campaigns hit 5.8% email, 14.4% warm LinkedIn)

• Discovery meetings booked: 5+ (2 already on calendar)

• Escalation: at least 1 contact brings a colleague or superior into a follow-up

👷 Challenges

• Mid-sized studio outreach produced significantly fewer replies than large studio outreach. The "KPI shift" framing that resonated with 500+ employee companies fell flat with 50-200 employee studios. The problem definition needs to be segment-specific rather than universal.

• The gap between campaign volume and meaningful conversations remains wide. Campaigns reached over 1,000 unique contacts across email and LinkedIn but converted to only 2 booked meetings. While cold outreach benchmarks hover around 1-3%, our earlier warm outreach at 5.8% (email) and 14.4% (LinkedIn) set higher internal expectations.

• Outreach tooling occasionally underdelivered on daily send quotas, hitting 40-42 emails on some days against a 60-email target. Root cause unclear after consulting the platform's support team. Mitigated by reducing planned volume and shifting overflow to LinkedIn.

• Competitor research has not yet translated into campaign strategy. Analysis of 10+ competitors confirmed most entered gaming through industry insiders or gaming-experienced investors, validating CogniScope's outsider challenge but not yet producing a replicable GTM playbook.

Every studio we've spoken to has dashboards, monitoring tools, and analytics teams. They know when something breaks. But even with all that analytics ability, the marketing team still complains about getting something actionable enough before the damage compounds. The pattern keeps surfacing: detection isn't the gap, speed of being able to take action is. Whether we can actually compress that window into something useful is what we're trying to find out.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Greetings from Team CogniScope 👋

The sprint where nothing shipped externally, but everything changed internally.

CogniScope is the role-based AI Sidekick suite for game studios. It removes the repetitive, operational work, so teams can ask, ideate, plan, and act from a single conversation. The system pulls information and pushes execution across the stack, helping game teams to focus on what matters most: building amazing games with their vision and creativity that players will love.

This was an infrastructure week. The backend completed its migration to Google Cloud Run, the evaluation system matured into a full experiment platform with cross-run statistics, and the team started a deliberate rethink of our outreach strategy. Two discovery meetings happened, Insight Partners reached out to build a relationship, and Teknopark extended our project deadline by five months.

All information shared in this newsletter is confidential, please do not share or forward.

📣 Asks

• We're exploring usage-triggered outreach, reaching studios at the moment their product data shows something interesting rather than cold. If you know anyone building or investing in signal-based B2B sales approaches, we'd love to compare notes.

• We have a meeting scheduled with Amazon Bedrock this week to explore their startup program. If you have experience with AWS startup credits or the Bedrock program specifically, a quick note on what to prioritize would be valuable.

• Warm introductions to Director+ stakeholders at mid-sized studios (10-200 people) running live-service games remain the highest-leverage help. Both discovery campaigns are active and we are scheduling meetings weekly.

• If you know anything about the tax treaty between UK-US corporations and even utilized it yourself, we would be happy to learn about it.

💼 Latest News

• The biggest change this week happened beneath the surface. Our backend has been deployed to Google Cloud Run as a standalone service, with every configurable parameter externalized into environment variables. The team can now change any setting instantly without redeploying the container, and a new live logging system enables real-time debugging straight from Cloud Run's console.

• On the AI side, the evaluation system took a major step forward. Anyone on the team can now launch experiments from a new dashboard, view aggregated statistics across multiple runs, and compare results with normalized scoring rubrics. This is our quality control infrastructure for the AI engine, and it matured well beyond the original goal this week.

• Insight Partners reached out proactively to build a relationship. Their team met with us on Tuesday, interested in vertical AI for gaming. A positive signal that the space is attracting investor attention.

• Two discovery meetings were completed this week from the Sudden KPI Shift campaign. The conversations confirmed our framing but energy was lower than expected, which is feeding into a broader rethink of how we approach outreach (more on that below).

• Teknopark extended our project deadline to August 2026, giving the team significant breathing room on the Turkish grant timeline.

🛠 Product

Cloud infrastructure:

• Backend deployed to Google Cloud Run as a new GCP service

• All parametric values previously hardcoded in the orchestration layer are now externalized as environment variables, editable instantly without redeployment

• Live container logging with verbose Cloud Run exports for real-time monitoring and fast debug filtering

• Admin panel prepared for monitoring and controlling data ingestion operations

AI evaluation system:

• Multi-run evaluation flow completed with a new experiment dashboard

• Statistical aggregation built for cross-run comparison, identified organically as a need during the work

• Evaluation rubric normalization for consistent scoring across experiments

• Health check cron system in progress for automated quality monitoring on a schedule

• Next: passing every test case with daily scheduling and multi-game support

🤝 Sales

With both campaigns running and early data flowing in, the team used this week to step back and question the approach itself.

• Outreach strategy rethink is the most interesting development. Rather than continuing pure cold volume, we are exploring a model where we wait for product usage signals (a DAU spike, a KPI anomaly) and reach out with relevant context. We are also building a warm relationship re-engagement list from previous meetings and action items, prioritized by potential and tagged in our CRM.

• Anomalies Validation (Email): 338 contacts, 640 delivered, 6 replies. 281 contacts still active in sequences.

• Proactivity Validation (Email): 502 contacts, 318 delivered, 2 replies. 416 contacts still active.

• LinkedIn channels: ~200 contacts across two LinkedIn campaigns, still ramping up.

• Combined pipeline: approximately 1,040 active contacts across all channels.

• The Riot relationship remains active with a complete deal status document and relationship map now in place.

📊 KPIs

Validation metrics framework:

• Meetings with budget-holding stakeholders

• Escalation depth (does the contact bring someone more senior?)

• Speed of problem pull within a sprint cycle

• Quality of objection patterns in replies

Sprint targets:

• Active outreach contacts: ~1,040 across 4 campaigns (2 email + 2 LinkedIn)

• Reply rate: 21 / 1,389 = ~1.5% cumulative. Campaigns still early, most contacts mid-sequence but early signals indicates a necessity of iteration.

• Discovery meetings this week: 2 (Sudden KPI Shift framing)

👷 Challenges

• Cold email reply rates are below target. The 1.5% cumulative rate across active email campaigns sits well below the 5.8% overall and 14.4% warm LinkedIn rates from our initial rounds. Campaigns are still early (most contacts mid-sequence), but the gap is driving the strategic pivot toward warmer, context-triggered engagement. Personal connection matters more than volume at this stage.

• Discovery conversations lacked urgency. Both meetings this week confirmed our problem framing but the prospects did not describe the pain as acute. This reinforces that cold outreach alone won't get us to the studios that need this most, and strengthens the case for signal-based timing.

• UK company closure paused. After consulting the accountant, we stopped our legal counsel from proceeding with the UK entity shutdown. Waiting on tax treaty clarity before making a final call.

One of our edges in previous years was being fast, quality builders. But building has never been easier or more accessible, so we need to focus more on the areas that AI cannot replicate well to make a bigger difference in the market. Our technical efforts should shift toward figuring out the "worthy" problem to fix in the industry before doing any more development. The world is going through a huge shift and we need to be ready for the next phase to survive and to lead. I wish everyone reading this (and myself) can step back and observe this shift at a larger picture.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Greetings from Team CogniScope 👋

Both campaigns scaling the same wall to see which one breaks through.

CogniScope is the role-based AI Sidekick suite for game studios. It removes the repetitive, operational work, so teams can ask, ideate, plan, and act from a single conversation. The system pulls information and pushes execution across the stack, helping game teams to focus on what matters most: building amazing games with their vision and creativity that players will love.

This week both discovery campaigns went live, our product architecture completed a major decoupling milestone, and a postmortem on the previous campaign gave us a sharper read on what language and channels actually create pull.

All information shared in this newsletter is confidential, please do not share or forward.

📣 Asks

• If you know Director+ stakeholders at mid-sized studios (10-200 people) running live-service games, warm introductions remain extremely valuable as both campaigns are now actively running.

• Introductions to product or analytics leaders at large studios are also welcome as the second campaign ramps up.

💼 Latest News

• Both discovery campaigns are now live and sending. With two LinkedIn accounts and four email inboxes running in parallel, we can reach approximately 1,120 contacts every three weeks. The email track sends an intro, a follow-up, and a break-up over three weeks. The LinkedIn track runs a connection request with a 6-day follow-up window. Contacts are split by channel based on whether we could validate their email.

• A full postmortem of the previous campaign (large studio senior PMs, KPI investigation framing) is complete. Across roughly 750 contacts, the overall reply rate landed at 5.8%. The strongest channel was warm LinkedIn outreach, which produced a 14.4% reply rate and generated the only booked meeting from the campaign. Cold LinkedIn and joint email-plus-LinkedIn sequences performed lower, confirming that personal connection matters more than volume at this stage.

• The reply analysis surfaced familiar themes: 23% described KPI investigation as annoying but manageable, root cause understanding and cross-team alignment came up repeatedly, and a new signal around the opportunity cost of investigation time is emerging from a handful of contacts. No one described the problem as an emergency they would chase down, which reinforces that our framing needs to sharpen around time cost and speed rather than pain alone.

• An investor reached out to us directly and wants to connect. We are scheduling an introductory call.

🛠 Product

Architecture Decoupling The frontend and backend are now fully separated from the orchestration layer. Chat calls, user data, and model requests all hit the database and backend directly. The only remaining dependency on the orchestration system is data ingestion and scheduling. This is a significant milestone: it means the core user experience no longer relies on a workflow engine, which improves speed, debuggability, and deployment flexibility.

Monitoring Dashboard A new internal dashboard is live, showing full logs for every run: each step in the agent flow, which model was used, token counts, cost per call, and timing. All agent workflows now write structured logs into a dedicated database schema, which means we can run queries like "show me every request that took longer than 60 seconds" without digging through workflow execution history. This replaces the old approach where debugging meant manually scrolling through execution logs in a third-party UI.

Next Steps The immediate product goals are migrating the remaining data ingestion and initialization processes to a cloud-native setup so we can fully retire the local development dependency, and building a multi-run evaluation flow that can execute the same test case many times and report averages, giving us statistical confidence rather than single-pass results.

🤝 Sales

Two campaigns are in motion, each testing a distinct problem:

• Campaign 1: KPI Root Cause Investigation (mid-sized studios) is live and sending. 508 director-level contacts are loaded into a multi-channel sequence. Early email metrics are still ramping (the system starts conservatively to protect deliverability), so meaningful signal will come in the second and third weeks.

• Campaign 2: Proactive Decision-Making (large studios) is in preparation. We have a 25,000-person filtered list from large gaming companies that needs further scoring and segmentation before loading. This week we are testing more cost-effective enrichment methods to get LinkedIn profiles evaluated at scale without burning through credits.

We have three meetings on the calendar for next week, including a Senior Product Manager at KaBam (via both a direct conversation and a one-pager forwarded to colleagues), a contact from Super Indie Games via Meet2Match, and a follow-up with a contact at Quali scheduled for March 22.

We are also re-engaging Riot with a follow-up this week and reaching back out to GameDev.ist now that GDC has wrapped, as they had indicated reconnecting after the conference.

The postmortem learnings are already shaping how we run these new campaigns: tighter sequences, cleaner channel separation (email-only or LinkedIn-only per contact, no mixing), and faster postmortem cycles so we can adjust messaging within weeks rather than at sprint boundaries.

📊 KPIs

We continue measuring validation strength through:

• Meetings with budget-holding stakeholders

• Escalation depth (does the contact bring someone more senior?)

• Speed of problem pull within a sprint cycle

• Quality of objection patterns in replies

3-week sprint targets (Campaign 1: mid-sized studios, 508 contacts):

• Meaningful response rate: 5%+ (previous campaign hit 5.8% overall, 14.4% on warm LinkedIn)

• Discovery meetings booked: 3-6

With Campaign 2 loading behind, we expect the combined data from both campaigns to give us a much clearer read on which framing and audience combination creates the strongest pull by end of month.

👷 Challenges

• Lead enrichment at scale remains a constraint. Our enrichment provider's AI integration is unreliable, and the workarounds add cost. We are testing alternative enrichment sources and evaluating whether running our own model calls against the provider's API could reduce per-lead cost by an order of magnitude. Getting this right unlocks the 25,000-person list for Campaign 2.

• CRM integration between our outreach tools and CRM needs tightening so that replies, messages, and contact data from both channels sync automatically rather than requiring manual exports. A planning session is scheduled this week to wire this up properly.

• UK subsidiary tax decision is still in progress. Our UK accountant flagged that UK subsidiary expenses may be deductible against US tax obligations under a tax treaty. We are coordinating between our UK accountant and a US tax specialist to evaluate this before making any decisions about the entity.

• Teknopark project extension has been submitted and is awaiting approval. The plan is to extend through October while we prepare the transition to an R&D center setup with a new project application.

• Website and discoverability remain on hold until we have clearer content direction. We explored technical SEO improvements (AI-readable metadata, LLM-friendly text files) but confirmed these only help if the base SEO is already working. Rebuilding the site is not blocking campaigns right now, but it will matter as inbound interest picks up.

The campaigns are live, the infrastructure beneath them is tighter than it has ever been, and the postmortem gave us a real feedback loop. Now we watch the data come in and adjust fast.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Greetings from Team CogniScope 👋

Finding your coordinates before making the next move. 🧭

CogniScope is the role-based AI Sidekick suite for game studios. It removes the repetitive, operational work, so teams can ask, ideate, plan, and act from a single conversation. The system pulls information and pushes execution across the stack, helping game teams to focus on what matters most: building amazing games with their vision and creativity that players will love.

This week was about converting weeks of groundwork into launch-ready assets. We went from 3,300 raw companies to 508 qualified director-level contacts at mid-sized live-service studios, our AI engine jumped from ~20% to 80-85% evaluation accuracy, and a new multi-channel campaign is loaded and ready to fire on Monday.

All information shared in this newsletter is confidential, please do not share or forward.

📣 Asks

• If you know Director+ stakeholders at mid-sized studios (10-200 people) running live-service games, warm introductions are extremely valuable right now as we launch our biggest discovery campaign yet.

• Introductions to product or analytics leaders at large studios are also welcome as we prepare a parallel campaign focused on helping teams shift from reactive firefighting to proactive decision-making.

💼 Latest News

• Our first campaign targeting senior PMs at large studios is wrapping up. 6% connection acceptance rate and 22% response rate from accepted connections. Early conversations, particularly with a contact at Coffee Stain, are producing strong validation signal and pushing us to explore adjacent problems around understanding the impact of actions taken on games.

• A targeted lead generation effort narrowed 3,300 companies down to 624 legitimate mid-sized live-service studios, then identified 508 director-level contacts with a custom scoring system. These are not generic leads: each profile was enriched with LinkedIn data, validated for active live-service involvement, and scored 0 to 5 on campaign fit. The top-scored contacts include heads of studios like Embark and senior leaders with deep experience across EA, Riot, and similar companies.

• An SEO analysis landed in our inbox this week noting CogniScope scores 25/100 on AI discoverability. We already knew this was a gap. Our current website setup limits control over crawlable content and metadata. Rebuilding it properly is now more prior with these mass outreach campaigns, so we own our discoverability.

🛠 Product

AI Engine Accuracy Jump
Our evaluation system that was averaging around 20/100 is now consistently hitting 80-85/100. This came from redesigning how the AI agents work together, improving prompts through structured debug cycles, and tightening how the system pulls and analyzes data. This is the first time our evaluations are being used to meaningfully improve outputs in a loop rather than just measuring them.

Infrastructure Hardening 
A major database optimization is underway that is already showing significant speed improvements on the queries our AI system relies on. The next milestone is deploying the backend to the cloud so we can move off local development and into a production-ready environment.

Proactive Actions Direction 
We had an honest internal conversation about the "proactive" product direction (predicting problems before they happen). The team agreed: the idea is strong, but we do not yet know enough to build something that would feel compelling in a demo. Rather than shipping a generic prediction UI, we will use the upcoming discovery conversations to learn what data studios actually have and what actionable output would look like in their world. For now, this stays as a draft, not a build target.

🤝 Sales

Two discovery campaigns are now in motion, each testing a distinct problem:

• Campaign 1: KPI Root Cause Investigation (mid-sized studios) - Launches Monday. The pitch: studios spend days of manual analysis to find why a KPI dropped. 508 director-level contacts are loaded into a multi-channel sequence that branches between LinkedIn and email based on connection status, with break-ups always going through email to preserve LinkedIn for future re-engagement.

• Campaign 2: Proactive Decision-Making (large studios) - In preparation. The pitch: product and analytics teams are stuck in permanent defense mode, always reacting to problems after they hit instead of getting ahead of them. The target list is currently being evaluated and scored using the same methodology as Campaign 1.

To support both tracks running simultaneously, we are adding a second LinkedIn account to our CRM, giving us two LinkedIn channels and five email accounts in parallel.

A post-mortem of the previous campaign (large studio senior PMs + KPI investigation framing) is also in progress to compare what people told us with what the product actually addresses and feed those learnings into both new campaigns.

📊 KPI’s

We continue measuring validation strength through:

• Meetings with budget-holding stakeholders

• Escalation depth (does the contact bring someone more senior?)

• Speed of problem pull within a sprint cycle

• Quality of objection patterns in replies

3-week sprint targets (Campaign 1: mid-sized studios, 508 contacts):

• Meaningful response rate: 5%+ (previous campaign hit 22% of accepted connections)

• Discovery meetings booked: 3-6

These are upper product contacts at studios with active live-service games. Even a small number of high-quality conversations from this pool will carry significant signal. With a large-studio campaign on proactive decision-making close behind, we expect the combined data points to meaningfully sharpen our product direction by end of month.

👷 Challenges

• CRM integration still needs cleanup before we can confidently sync data across our outreach tools, LinkedIn, and CRM without duplicates or overwrites. This is scheduled as a planning session this week.

• Teknopark logistics: We are entering conversations about extending the project and transitioning from incubator status to an R&D center, which opens different tax and operational benefits. This adds short-term overhead but is strategically important.

• UK subsidiary decision: We discovered through our UK accountant that keeping the UK entity open may offer tax treaty benefits, where UK subsidiary expenses can offset US tax obligations. We are coordinating between our UK accountant, US accountant, and legal counsel to make the right call, rather than rushing to close.

The machine is more precise than it has ever been. Lead quality is higher, evaluation scores are real, and the campaign infrastructure is battle-tested. Now we convert structural readiness into live conversations.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope

Greetings from Team CogniScope 👋

Chloe escaped San Francisco and showed up in Istanbul. Some of you already know how that goes :)

CogniScope is the role-based AI Sidekick suite for game studios. It removes the repetitive, operational work, so teams can ask, ideate, plan, and act from a single conversation. The system pulls information and pushes execution across the stack, helping game teams to focus on what matters most: building amazing games with their vision and creativity that players will love.

This week was less about visible motion and more about compression. We focused on tightening internal systems so that every unit of effort produces more leverage. The result: fewer moving parts, clearer visibility, stronger control.

All information shared in this newsletter is confidential, please do not share or forward.

📣 Asks

• Introductions to industry event speakers would be hugely helpful as we prepare talks and roundtables.

• If you can introduce us to Director+ level product stakeholders at mid-sized studios, that would be highly valuable as we expand beyond top-tier publishers.

💼 Latest News

• Outbound continues across channels:

  • ~800 leads currently in the active system

  • 300 leads poked in this cycle

  • 6 replies so far, 1 negative

  • 3 meetings set via Meet2Match with senior stakeholders

  • 1 meeting set via LinkedIn

• LinkedIn conversations are surfacing adjacent problem areas beyond our current framing. These insights will directly inform the next discovery sprint.

• Lead segmentation is now structured. We began with senior product leaders at large studios and are now expanding to Director-level stakeholders at mid-sized studios. Campaigns are running automatically, with iteration happening week to week based on signal rather than assumptions.

• Separately, we continue monitoring competitive activity around the problem space that initially brought us into Techstars. Several new startups are now attacking that area. While we pivoted away due to lack of urgency from the market, this activity reinforces that our early exploration gave us a structured understanding of that space. We view it as validated research now embedded in our arsenal rather than unfinished business.

🛠 Product

This was a foundational product week.

Agentic System Compression
The agentic data scientist has been replicated in a fully enhanced environment with significantly improved control. Key upgrades:

• Major latency improvement

• Accuracy improvement

• Structured logging at every step

• Refined evaluation rubric for targeted performance improvements

• Clear cost and token visibility per call

The system is now in internal confidence-building mode. It is not yet externally positioned as production-ready, but the architecture is stabilizing in a way that materially increases our control and debuggability.

The focus is precision over expansion.

Clickable Mock-Up
A live, clickable mock-up for the current problem sprint is ready. It is intentionally minimal and static, built to:

• Explore the solution space while building

• Apply real-life scenarios heard from leads

• Make abstract conversations tangible

These prototypes are proving valuable not just for demos, but as thinking tools.

🧠 Operations

Operational compression was the dominant theme this week.

CRM and lead infrastructure:

• Semi-automated internal tool built to import LinkedIn contacts, companies, and conversation history directly into our CRM

• Meeting note tracking partially automated across Attio, Fireflies, and internal logs

• Lead segmentation clarified and active

Finance and compliance:

• US and UK CPA/tax services finalized

• Filings prepaid and structured

• Banking security updates resolved

• Ongoing education to maintain long-term control over cross-border obligations

Reducing administrative drag is already increasing speed across sales and product loops.

📊 KPI’s

We continue measuring validation strength through:

• Meetings with budget-holding stakeholders

• Escalation depth

• Speed of problem pull

• Quality of objection patterns emerging in replies

The current signal is still early-stage discovery. The goal remains increasing urgency clarity, not volume for its own sake.

👷 Challenges

• Translating surfaced adjacent problems into a coherent narrative without diluting focus. Discovery breadth is expanding; discipline must expand with it.

• Teknopark extension process begins this week. While manageable, it adds short-term overhead.

The machine is tighter than it was last week. Fewer leaks, more observability, more control. The next step is converting that structural strength into sharper external pull.

If you'd like to quickly catch up let's schedule 15-minute call here!

All information above is confidential, please do not share or forward.

Written with ❤️,
Team CogniScope