The story so far.
Where lucix came from, what we killed and why, and where we stand today. Weekly updates follow below, newest first.
Nick & Ryan · updated weekly · click to read the whole arc
The story so far.
Where lucix came from, what we killed and why, and where we stand today. Weekly updates follow below, newest first.
Nick & Ryan · updated weekly · click to read the whole arc
lucix exists because of one sentence, and everything since has been in service of it: getting a group of people to actually agree is one of the most common, most expensive problems in business — and almost nobody has good tooling for it. Neither of us arrived at that as a market-sizing exercise. We lived it, from two different sides, for years before we ever thought about building a company around it.
I spent over a decade around Imagen, a video asset management platform, most of it in sales and revenue leadership. In between, I ran sales for Diversified, where I watched good enterprise deals stall — not on price, not on product, but on cross-functional buy-in that never quite landed. I went back to Imagen as CRO for the Americas, and shortly after, we were acquired by Thomson Reuters. The integration surfaced the same problem from the inside: decisions made without real visibility or agreement across a much bigger organization made what should have been a straightforward acquisition far harder than it needed to be. Trying to fix that from inside a company far larger than the one I'd help build, and not fully succeeding, is a real part of why I left.
Ryan Burrus joined as co-founder in January 2026, running go-to-market from day one. His background is enterprise B2B product marketing — at BlackBerry he ran go-to-market for a global business unit on a $1.2M budget and built account-based marketing that generated $9.9M in pipeline, on top of six years at Verizon before that. Nearly everything below on real market interest is Ryan, largely by hand, with no marketing budget at all.
What follows is the honest version — including the parts that didn't work.
Three ideas, one thesis, each one killed on purpose
lucix has been built three times. That's not indecision — it's discipline. Each time, we found out something the market was telling us clearly, and instead of forcing the existing idea to fit, we stopped and rebuilt around what we'd actually learned.
The idea, before it had a name
Two months between roles. An idea about visualizing organizational decisions, validated informally, then rebuilt twice in three weeks as the real thesis sharpened. By mid-January, the name was lucix.
A production platform, a real go-to-market motion, and an honest dead end
Five months, 291 commits, a full multi-tenant application with real security architecture, deployed live. We ran 25–30 structured discovery interviews, built an active outbound pipeline, and demoed the product in person to enterprise prospects. The market's answer was consistent: the mechanism itself, provoking a reaction from stakeholders, wasn't the right one. So in June, we stopped shipping it.
A prototype, deliberately, and the pivot that mattered
Before writing a line of new code, we ran a structured adversarial exercise against our own new idea: steelmanning the alternatives, running a premortem, then steelmanning ourselves. Only once that came back “genuinely novel, real business case” did we build — and we built a prototype on purpose, deliberately narrow, and held it in prototype mode past the point where graduating would have felt natural. It modelled a single decision as the unit of everything: a scoping choice we made deliberately and recorded as temporary, because it was the cheapest structure that could answer the question we were actually asking. Three weeks in, while scoping how the product would sit inside Slack, we ran the query that cashed it in — all 42 decisions in the prototype were single-origin, exactly as designed, and nothing in the schema carried the same person across two of them. That was the prototype doing its job. The constraint we had knowingly accepted sat precisely where our real buyer turned out to live: the Chief of Staff, the person who owns getting a decision made, not the decision itself. Its answer in hand, we retired it and priced the rebuild.
The real build
We rebuilt the prototype into a real product codebase: a tenant-per-customer data model with row-level access control, an independent team adversarially reviewing our own architecture before anything shipped, a real production database. Yesterday, August 3rd, I logged in as the first real user and pulled my own real record through the entire live stack: identity, permissions, database, all of it, for the first time in this product's history.
Everyone else uses AI to replace the human in the loop. We use it to make the human in the loop more powerful.