Meeting Intelligence OS
Your meetings make decisions. Your tools only remember the words.
THE PROBLEM
Every organization runs on decisions made in rooms — and loses most of them almost immediately. The transcript survives. The recording survives. What doesn't survive is the thing that actually mattered: what the team decided, what alternatives it weighed, and why it chose the path it did.
The cost stays invisible until it isn't. New hires re-open settled questions. Teams discover months in that they've been working from different versions of the same agreement. Leaders defend choices made under conditions nobody wrote down. Institutional memory walks out the door with whoever leaves. Most meeting tools make this worse, not better — they give you more to search through, not more clarity.
THE OUTCOME
Organizations that work this way accumulate clarity instead of losing it. Decisions become traceable.
Onboarding shifts from absorbing output to understanding reasoning. Cross-functional teams operate from one shared record rather than four partial ones. The question "why did we decide this, and what did we consider?" finally has an answer that doesn't depend on who's still in the building.
Where this stands
We're rolling these systems out in stages, deliberately. If decision memory is a problem you're already trying to solve, we'd like to talk with you early.
[Book a demo] · [Join the early-access list]
OUR APPROACH
We treat a meeting as a unit of organizational learning, not an event to be transcribed and forgotten. That means capturing three layers, not one.
The first layer is what was said — accurate, searchable, the table stakes. The second is what was actually decided, separated cleanly from the discussion that produced it. The third is the reasoning: the alternatives considered, the assumptions made, the dissent raised and either resolved or noted.
Those layers are designed to be queried over time, so the knowledge compounds instead of scattering. And the system is built to assist the people in the room rather than replace their judgment — surfacing what matters, doing the clerical work, and leaving the decisions to the humans accountable for them. Data is handled with the discipline our team carried over from years in fintech and banking: held in trust, never treated carelessly.
THE PROBLEM
For as long as games have been made, creators have faced the same trade-off. Choose the powerful tools and spend months learning them before anything is playable. Choose the accessible tools and accept a ceiling somewhere short of your ambition. Most people pick one, hit the wall, and either grind through it or give up.
The industry's usual answer has been to bolt assistive AI onto the tools that already exist — generators for assets, copilots for scripts. Useful, but it only makes the old workflow cheaper. It doesn't change the shape of the work, and it doesn't move the wall.
OUR APPROACH
We start from a different question: not "what can we automate inside the existing pipeline," but "what should the pipeline be when it can be different."
That means designing creation around what matters and removing the overhead around what doesn't. The friction should sit in the creative decisions — what kind of game you're building, how it should feel — not in the technical translation of those decisions into something that runs. Iteration speed becomes the thing we optimize for, because most ideas die in the gap between "I can describe it" and "I can play it." Close that gap and you change who gets to make games at all.
Our team's background spans digital games, data science, and design, and we build with the same principles that run through everything we do: systems that augment the creator rather than replace them, and that treat the people using them as the ones in charge.
THE OUTCOME
Creators get from idea to playable build without the technical tax that usually stands in the way. The barrier to entry drops without the ceiling dropping with it. The line between playing and creating softens, so the tools that shape an experience start to feel continuous with the tools that build it. Power and accessibility stop being opposites.
Where this stands
Most of this work is still in the lab, and what's public is a fraction of it — by design. If you're a creator, studio, or partner already building as if this future has arrived, we want to know who you are.
[Join the early-access list] · [Book a demo]
Gaming & Creation
Research & Analytics
Half of your most important data never reaches a dashboard.
THE PROBLEM
The signals that get tracked, charted, and reported are the ones that were easy to count. The signals that often matter most — what a customer actually said in an interview, the hesitation in a survey response, the comment buried in an open-text field — sit in a folder somewhere, or never get analyzed at all.
This isn't a failure of intent. It's a cost problem. Quantitative data scales cheaply; a pipeline and a dashboard and you're done. Qualitative data has always scaled in human-time — transcribing, coding, interpreting, rewriting. The cost per insight was high enough that most organizations quietly stopped asking the qualitative questions. The result is a generation of analytics built for the half of the picture that's measurable, not necessarily the half that matters.
OUR APPROACH
We build analytics systems that treat qualitative and quantitative data as one connected fabric, rather than two separate disciplines that meet once a quarter in a slide deck.
The shift is in economics, not magic. The bottleneck steps in qualitative work — transcription, initial coding, theme clustering, cross-source synthesis — compress dramatically when the right systems do the grunt work. That isn't a productivity story; the productivity is a side effect. The real change is what becomes affordable to study. Questions that used to be quarterly investigations become continuous. The rigor of qualitative coding survives the move to scale, and the speed of quantitative reporting extends to evidence that was previously too expensive to gather.
This is the part of our work closest to where the team started — years across research, analytics, data science, and behavior-change communication, much of it in African markets where the qualitative signal isn't a nice-to-have but the only way to understand what's actually happening. Human judgment stays at the center; the system removes the drudgery that used to crowd it out.
THE OUTCOME
Teams stop choosing between what they can measure and what they believe is true. Dashboards combine what happened with what people said about it. The questions you can afford to ask broaden, and they update as your population does. Decisions get made on a fuller picture — and the bar for what counts as enough evidence rises, because gathering it no longer costs what it used to.
Where this stands
We're rolling these systems out in stages. If your research function has been trying to bridge the measurable and the meaningful, we'd like to talk with you early.
[Book a demo] · [Join the early-access list]
Marketing & AI-CRM
Most marketing systems describe the past. We build for the present.
The Problem
Marketing teams have more data than ever and less clarity about what to do next. Dashboards report opens, clicks, conversion windows, churn cohorts — every metric the team agreed mattered six months ago, displayed in confident colors. Almost all of it describes what already happened.
The reason isn't capability; it's design. Most platforms are built around segments — static buckets a team defines once and revisits quarterly. Real customer behavior doesn't sit in buckets. A high-intent buyer this week is a lapsed lead next month and a referral source next quarter. A system that updates its picture once a campaign is, in effect, flying with last year's weather report. And as generative tools flood the market, a second problem creeps in: everything a brand publishes starts to sound like everything every other brand publishes.
Our Approach
We build marketing and CRM systems that work in the present tense — continuous rather than quarterly, prescriptive rather than merely descriptive.
The unit of analysis shifts from the segment to the signal: a system that updates its understanding of a customer on every interaction recognizes movement instead of forcing people into predefined buckets. The marketer's role shifts from operator to editor — campaigns get steered as they run, not authored once and sent. And brand voice is treated as a learned property of the company, encoded in the brand's own best work and the feedback loop around it, not as a paragraph pasted into a prompt. The aim is generative marketing that makes a brand more itself over time, not less.
This sits squarely in our team's experience — fintech, banking, media, and behavior-change communication, where the discipline of handling customer data carefully and understanding why people actually act the way they do isn't optional. Data is held in trust, and human judgment stays in the loop.
The Outcome
Teams stop reacting to last quarter and start acting on this moment. Decisions arrive with a reason attached, not just a metric. Volume sent stops being the measure of success; what matters is whether the system produces the decisions a sharp team would have made on its best day, faster. And the brand keeps sounding like the brand, even as output scales.
Where this stands
We're rolling these systems out in stages. If this is the direction you've been trying to move your marketing in, we'd like to talk with you early.
[Book a demo] · [Join the early-access list]