Where most engagements start
Four weeks to find out, at a price agreed up front
Scope, price and acceptance criteria signed before we start. At the end you have working software and an evaluation report, or a written case for not going further. Both are useful; only one is expensive to discover later.
AI Pilot
The situation
Before we start
Budget built on a guess
Approving a program without a costed plan means padding the number or discovering the real one halfway through. Neither survives a finance review.
Nobody checked the data
Most stalled projects fail on data assumed to exist, or existing without the labels that make it useful.
Vendors cannot say no
A firm paid to build is badly placed to advise against building. A fixed fee that can end in a recommendation to stop removes that conflict.
What working means here
Signed first. Measured after.
The value of each one is set with you during scoping, from your data and your line. What does not change is that they are written down first and published against afterwards, pass or fail.
- 01
Acceptance criteria met
The numbers signed in week one, reported against in week four
- 02
Evaluation method documented
Held-out set, procedure and result, so anyone can repeat it
- 03
Costed path
Phases, schedule and price bands for production, or the reasoning for stopping
How it is assembled
Stage by stage
Each one ends in something you keep.
- Step 01
Week 1 — scope and sign
Workshops, data review, and acceptance criteria agreed in writing. Nothing is built until both sides have signed what success means.
- Step 02
Week 2–3 — build
Working software against a real slice of the problem, on your data, reviewed with your team mid-way.
- Step 03
Week 4 — evaluate
Tested against the criteria from week one, with the method and the held-out set documented.
- Step 04
Readout
The result, a costed path to production, and a clear recommendation. Including do-not-proceed, which we have given before.
Scope
What the work includes
- Signed acceptance criteria before any code is written
- Workshops with the people who do the work today
- A site survey where the work is physical
- Working software against a real slice of your problem
- Evaluation against a held-out set, with the method documented
- Reference architecture and integration outline
- A costed, phased path to production with a schedule
- Readout with your team and your budget holder
Deliverables
What you keep
- 01Signed acceptance criteria
- 02Working software
- 03Evaluation report against the criteria
- 04Reference architecture
- 05Costed path to production
- 06Recommendation and readout
Manufacturing · Vision
Surface defect inspection, decided at the camera in 12 ms
Final inspection was manual and sampled, so surface defects were being found by the customer rather than on the floor. We surveyed the cell, captured production across three shifts, and commissioned an edge vision station that inspects every part and drives the existing reject gate.
Read the engagement- 12 ms
- decision latency at the camera
- 99.4%
- detection rate on hold-out set
- 100%
- of parts inspected, up from sampled
Start
How AI Pilot begins
$28K–60K, fixed price, four weeks
Where an engagement lands inside its band is set by data readiness, integration depth, the reliability bar, evaluation burden, deployment constraints and usage economics. All six, explained.
No proposal, no pipeline stage. The next step is the engagement itself.
Scope a pilotFAQ
Asked before signing
The pilot is priced as its own piece of work and delivers its own artefacts, so it is not discounted afterwards. What it removes is the guesswork in the production quote, which is where the real money is.
Then you have that in writing with the reasoning, for a known fixed cost, instead of finding out in month five. It has happened and we will say it plainly.
Where in the $28K–60K band it lands is set by data readiness, integration depth and whether a site visit and hardware are involved. The number is fixed once scope is signed.
Whoever does the work today, someone who owns the systems it will touch, and the budget holder. Around eight to twelve hours of their time across four weeks.
Often paired with
Where this fits with the rest
- 01Core AI
End-to-End AI Development
First workshop to an operated system, one accountable team. Agents, retrieval, custom models.
- 02Physical AI
Edge AI & Computer Vision
Detection, inspection, tracking and safety on-device. Milliseconds, and no network dependency.
- 03Start here
Vision Feasibility Check
Send 30–50 sample images. Get a written memo: accuracy band, hardware class, risks, go or no-go.