Entry engagement — start here
AI Adoption Sprint
One high-value AI use case, prototyped and evaluated honestly.
- Duration
- 3 weeks
- Format
- Remote, workshop plus working prototype
- Fee
- Fixed fee, from $12,000
If you go on to a build engagement with us, this fee is credited to it in full — so a confident “yes” to the work costs you nothing extra for the discovery.
This is for you if
- There is pressure to 'do something with AI' and no agreed definition of what would count as success.
- You have a specific workflow — document handling, search, triage, summarization — that looks like a candidate.
- You need to move without creating a data-protection or client-confidentiality problem.
This is not for you if
- You want a strategy document rather than something working.
- The data the use case depends on does not exist yet, or cannot be shared with us in any form.
The problem it solves
Most AI initiatives stall in one of two ways. They stay slideware, because nobody ever built the thing. Or they ship a demo that impresses in a meeting and quietly fails in production, because nobody defined what accuracy would be acceptable before they started.
The useful discipline is unglamorous: pick one measurable use case, decide the acceptance threshold in advance, build the smallest real version, and evaluate it honestly against that threshold.
How it runs
What we actually do
-
Week 1 — Inventory and choose
Workshop to surface candidate use cases, scored on business value, data readiness, risk and feasibility. We pick one. We also name the ones that do not need AI at all.
-
Week 1 — Set the bar
Define the measurable acceptance criteria, the evaluation set, the privacy and security constraints, and what the fallback is when the model is unavailable or unsure.
-
Week 2 — Prototype
Build the smallest version that exercises the real workflow on real data, with human review in the loop.
-
Week 3 — Evaluate and decide
Accuracy, cost per transaction, latency, failure modes and usability measured against the threshold set in week one. Then a clear recommendation: proceed, revise, or stop.
Deliverables
What you receive
Documents you keep, whether or not you continue with us.
- Business use-case inventory, scored and ranked
- Data-readiness evaluation for the chosen use case
- Security, privacy and confidentiality assessment
- A working prototype exercising the real workflow
- Evaluation results against acceptance criteria agreed in advance
- Cost model at realistic production volume
- Build-versus-buy recommendation
- Internal AI usage and data-handling policy for your developers
- A prioritized plan for the next pilot, or a documented reason to stop
What happens afterwards
A successful sprint usually leads to a production implementation of the same use case, and often to a second sprint on the next candidate. Roughly a third conclude that deterministic software solves the problem better — which saves considerably more than the sprint costs.
Questions we are usually asked
What if the honest answer is that AI is the wrong tool?
Then that is the deliverable, with the reasoning and the cheaper alternative. It is the most valuable outcome we can produce for some clients, and we would rather say it in week three than after you have spent a year on it.
Will our data be used to train someone's model?
No. Data handling, retention and vendor terms are part of the week-one assessment, and we design for data minimization and access control from the start.
Can you help our developers use AI tools rather than build AI features?
Yes — that is a different engagement, and often the higher-return one. See Development Team Services.
Do you have a preferred model or vendor?
No. The choice follows the accuracy, cost, latency and data-residency constraints of your use case, and we will show you the comparison.