AI Sprint
Two weeks from question to working prototype — audit, roadmap, and an AI system you can put in front of users.
- Format
- FIXED SCOPE — 2 WEEKS
- Price
- €3–6K, fixed
- Terms
- Fee credited against the build
- Ladder
- Rung 01 of 05
00 — Best for
- You tried an AI pilot and it never left the demo.
- You know AI should be doing some of your work, but not which part.
- You need a working prototype before committing a real budget.
01 — Why a sprint
Most AI projects fail before they start — scoped by a slide deck, priced by guesswork, built against data nobody looked at. The sprint replaces that with two weeks of real work: we take one process, your actual data, and build a prototype that either proves the case or kills it cheaply.
Either outcome is worth the fee. A working prototype becomes the spec for the build. A dead end found in two weeks costs a fraction of one found in month four.
02 — How it runs
Week one is the audit: your data, your tools, the process and its real cost. Week two is the build: a prototype connected to a sample of your data, demoed live, warts included.
You end the sprint with the prototype, a production roadmap, and a fixed quote for the full build. Proceed within 30 days and the entire sprint fee is credited against it.
03 — What you get
- Audit of your data, tools and the process worth automating
- A working prototype on your data — not slides
- Production roadmap with architecture and model choices
- Fixed quote for the build, valid 30 days
- Full credit of the sprint fee if you proceed within 30 days
04 — Straight answers
What if the prototype shows AI is the wrong tool?
Then the sprint saved you the build. You keep the audit and the roadmap — including the recommendation not to build. That happens, and we say so.
Whose models and infrastructure?
Yours. We build on your accounts — model providers, cloud, data — so nothing is held hostage. You own everything the sprint produces.
Do you need our data before the sprint starts?
A sample, yes — enough to prototype against. Week one includes getting access sorted, and a DPA is part of the paperwork when personal data is involved.
Sound like your next step?
Start a project