Working together
Three ways to engage.
Most agencies push whichever model bills most. These are described plainly so you can pick the one that actually fits, including the cases where the answer is none of them.
Fixed-scope project
A defined outcome, a fixed price, and explicit exit criteria. Best when you know what you need built.
We scope in phases with a decision point between each. You can stop after any phase and keep everything produced so far — code, documentation, evaluation harness, infrastructure definitions. This is the right model for a first engagement because it caps your risk while we prove we are worth continuing with.
Embedded pod
A small senior team working inside your process, on your board, for a defined period. Best when requirements are still moving.
Typically two to four engineers with a technical lead, working in your sprints and reporting to your people. Suited to AI product work where the specification genuinely cannot be pinned down in advance because you are still learning what the model can do. Priced monthly, minimum three months, because anything shorter is mostly onboarding.
Ongoing retainer
Continuous work on systems already in production: evaluation, monitoring, cost tuning and incremental features.
AI systems drift. Models get deprecated, indexes go stale, providers change pricing, and the accuracy you measured in March is not the accuracy you have in September. A retainer covers that maintenance plus a defined amount of new work. We will not sell you one for a system that does not need it.
When none of these fit
If your problem is solved by an off-the-shelf tool, we will tell you which one and take no fee.
This happens more often than the industry admits. It costs us a project and earns us a referral, and over a decade that trade has been strongly positive.
Work out which model fits
Tell us what you are trying to build. We will tell you honestly whether we are the right team for it, and what it would realistically take.