How we work
Execution is a system, not a sales pitch.
Below is the operating model in enough detail to check it: the four stages of a client journey, and the delivery roles and pricing behind them.
Written down so you can check it.
Describing the operating model in enough detail that a client can check it, and then being held to that description on a real engagement, is the standard this page is written to.
So this page is more specific than marketing needs it to be, on purpose: which stage produces which artifact, and which decisions stay with a named person.
Every claim below names the artifact it produces, so each one can be checked.
Client journey
The client journey, stage by stage.
Understand, design, build, evolve. Each stage describes what we do together, and each outcome is the state it leaves you in, whether or not you continue to the next.
Four stages on one line. Select a stage to read what it produces, and what you are left holding when it ends.
Each stage ends with something you keep: a shared context, a plan, a system in production, or a team that can run it.
Where you enter depends on what you already have.
Not everyone starts at stage one. An organization that already has clarity needs execution, and one that has been building for a year may need an honest read on whether the path still holds.
- Understand or Design
- Advisory
- Design or Build
- Engineering
- Build, with parts already solved
- Engineering, with its accelerators
- Prove it first, then decide
- Phase Zero
Phase Zero is the exception to the shape. It compresses understand, design, and build into one small engagement on a single process, so the proof arrives before you commit to a stage at all. What it leaves you with is a baseline, a working solution, and a roadmap, which is the strongest position to enter any of the four stages from.
Tell us which stage you are in, and we start there.
Delivery model
How the work gets done.
AI agents run across every major role in an engagement, from discovery through deployment, with a named senior engineer accountable for judgment and outcomes. Test coverage, documentation, and the reasoning behind each decision are captured as the work happens, before the schedule can push them out. That is what lets a small team commit to a price and carry the risk of its own estimate.
Engagement model
We price for value.
The goal is not to sell more hours. Selling hours means our incentive improves when the work takes longer, and we would rather not build a business on that.
Large time and materials programs are getting harder to justify, and reasonably so. They place the risk of overrun entirely on the buyer, and they reward the seller for the thing the buyer least wants.
At the same time, buyers expect AI to create real efficiency. If a firm claims AI has transformed its delivery and still bills the same hourly way it did five years ago, one of those two things is not true.
We price against the result, and the shapes below are how that works in practice.
The shapes the work takes.
Which one fits depends on how much is already known. The first is built to come first.
A fixed-scope pilot
Phase Zero: one named process, about a month, and a fixed fee agreed before we start. You keep the baseline, the working pilot, and the roadmap whether or not you continue.
An embedded team
A small senior team, with its agents, working inside your organization beside your own people on a live backlog. The scope moves as the work does, and the price follows the outcomes.
An outcome-priced build
A defined system delivered end to end, for a price tied to what it has to do in production. The risk of the estimate is ours, and the baseline a Phase Zero produces is what that price is argued from.
The delivery model is what makes the commercial model possible.
When delivery speed is set by headcount, the only honest thing to sell is time, and the risk of everything taking longer sits with you. Our delivery runs AI agents across every major role in an engagement, which compresses the work enough that we can carry the risk of an estimate ourselves.
Reuse compounds the same effect. Each engagement produces methods and patterns that lower the cost of the next one, so a meaningful share of the work is not being invented on your budget.
It is also why the first engagement is the small one. Phase Zero produces the baseline a value price has to be argued from, and until that number exists the risk of an estimate is ours to carry.
Start with a pilot.
The smallest engagement, and the one that sets the baseline.
A pilot running beside production on one process you name, with a roadmap for what comes after it. The duration and the fee are on the Phase Zero page.
Read it, then test it.
Everything on this page can be checked in a first conversation. Start from the stage you are in.
Start a conversation