How we work · 07
Built with you,
not for you.
With equity on the table.
A new AI product built as a venture rather than as a project. Part of the consideration is taken in equity — which means we carry the same downside you do if the thing finds no market. Scope is set by a commercial thesis agreed before any code.
You hold: the venture, the written commercial thesis it was built against and the same build artefacts as any other engagement: harness, monitoring, ADRs, runbooks.
Why equity changes the engagement
We are exposed to whether it works, not just to whether it ships.
A project ends when the deliverable is accepted. A venture ends when it has a market, or when it is honestly wound down. Those are different incentives, and taking part of the consideration in equity is how we put ourselves on the second one. It also changes what we will argue with you about: a services firm has no reason to challenge your commercial thesis, and we have every reason to.
So the thesis comes first, in writing, before code. Who the buyer is. What they do instead today. What would have to be true for this to beat that. We have declined ventures at this stage, and the fee we did not earn was cheaper for both sides than the build we would otherwise have delivered.
This is the studio side of the business, funded by the consulting and build work. It is available to partners who want that capacity pointed at their sector — the ones who know a market properly and need the technical half built by people who will still be exposed in year two.
The best things we build outgrow us
A product that still needs the lab has not finished. We build ventures to stand alone — spun out, handed over, or opened up — and we count that as the win. Release is the rule; retention is the exception.
How it runs
Same discipline. Different ownership, term and exposure.
Delivery runs on the same governed workforce and the same definition of done as every other engagement. What differs is who owns it and who carries the risk.
- DurationScoped per venture
- ConsiderationFee and equity, agreed at the outset
- GateA written commercial thesis before a line of code
- DeliveryThe same governed workforce, the same definition of done
- ExitAn agreed wind-down path, written before we start
Write the commercial thesis
Who buys this, what they do instead today, what would have to be true for this to beat it, and what would falsify the whole idea. Written down, and argued rather than agreed.
Settle ownership before anything is built
Equity, IP, and what each side holds if it succeeds or is wound down. Agreed at the outset — never negotiated once code exists and leverage has shifted.
Build it the ordinary way
Evals before prompt code, failure-mode analysis, review gate, integration gate, human decision. A venture is not an excuse to lower the definition of done.
Test the thesis, not the build
The measure is whether the market responds, not whether the sprint completed. Where the thesis is failing we would rather say so early, and the equity is what makes that conversation honest.
Stand it up, or wind it down
Spun out, handed over or opened up when it can stand alone. Or wound down along the path agreed at the start, which is the reason that path is agreed at the start.
The two lists that matter
What you hold, and what we will not do.
Both are in the engagement letter before you sign it. The second list is the one worth reading twice — it is where most disappointment in this market actually comes from.
What you hold at the end
- The venture — its code, evals, documentation and IP — on the terms agreed at the outset
- The written commercial thesis it was built against
- The same build artefacts as any other engagement: harness, monitoring, ADRs, runbooks
- The ownership and wind-down terms, settled before work started
- A partner exposed to the outcome, not just to the invoice
What we won’t do
- Ventures without a written commercial thesis
- Equity arrangements agreed after work has started
- Ventures where we would hold IP you need in order to operate
- Building a competitor to a current client
- Ventures with no agreed wind-down path
Where this comes from
The lab exists to make the thing the market cannot buy.
If it can be bought, we buy it, and we spend our people on the part nobody has solved. Every venture in this lab has someone whose name is on it who would be building it anyway — and we would rather run one short-handed than hand it to someone doing it for the rate.
The six demands →Test it first
Most ventures should start as a Feasibility Sprint. Two to three weeks and a fixed fee is a cheap way to find out whether the thesis survives contact with a working system.
Have a market we should build into?
Thirty minutes with a founder. Bring the thesis, not the feature list — and expect us to argue with it before we agree to build anything.
Fee and equity agreed at the outset · Commercial thesis first · NDA available