← All services
Service · 08

Stalled AI Prototype Review

Your AI prototype impressed everyone at the demo and has been sitting still ever since. In 5 working days I test it on your data and give you a verdict: ready for a Sprint, fix first, or kill it. With the number we will measure from.

Who it’s for: For founders and CTOs with an AI prototype, demo or pilot that never made it into daily work, who need to decide whether to keep investing in it.

If the system already runs in production and fails, the right service is the AI Rescue. If nothing has been built yet, start with the Opportunity Map.

Message me on LinkedIn
time
5 working days
format
code and data review
outcome
verdict + baseline metric
price
from €2,200 net
( 01 )

The demo impressed. Then nothing.

The prototype took a few weeks to build and the board has seen it, yet for a quarter nobody has been able to say whether it will become a product. The team has other priorities, ownership has blurred, and every discussion ends with “it needs a bit more work”. After five days you know whether the prototype is worth a Sprint, what needs fixing first, or whether to kill it.

( 02 )

What you get

  • A review of the prototype: code, prompts, data and the process it is meant to serve
  • A test on a few dozen real cases from your company, and its result
  • A list of what blocks rollout: data, quality, integrations, missing ownership
  • A baseline metric: the one number a Sprint should improve
  • A two-page verdict: ready for a Sprint, fix first, or kill it
  • If ready: the scope and price of the Sprint
( 03 )

How it runs

01Day 1A call with the process owner and a review of the prototype: code, data, goal.
02Days 2–⁠4A test on real cases, measurement and the list of blockers.
03Day 5Verdict and review: what comes next and what it costs.
( 04 )

What you need on your side

The review works if you have:

access to the prototype: the code or a working version
a sample of real data or cases
the process owner for an hour on day 1 and for the review on day 5
readiness for a “kill” verdict
( 05 )

Price

from €2,200net

The starting price covers one prototype and one data sample, tested in five working days. Several prototypes or many data sources are quoted separately, before we start.

  • order a Sprint for the same process within 30 days of the review and the full fee comes off its price
  • a “kill” verdict is included: you get the reasons, with numbers

Tell me what the prototype does and what data it should run on. Within 1–⁠2 business days I’ll reply with a price and a start date.

As of: November 2026

( 06 )

Before you write.

How is this different from the AI Rescue?

The Rescue fixes a system that runs in production and fails. The review assesses a prototype that never went into use and decides whether it is worth rolling out at all.

What if the verdict is “kill it”?

You get the reasons with numbers, and a list of what would have to change for the idea to come back. That costs less than another quarter of attempts.

Do I have to order a Sprint afterwards?

No. The verdict, the test set and the metric are yours. If you order a Sprint for this process within 30 days, the full fee of the review comes off its price.

How much of my team’s time does it take?

About three hours: a call at the start, answers to questions along the way, and the review of the result.

( 07 )

What’s next

A “ready” verdict leads to the Agentic Workflow Sprint: the test set and metric from the review become its first week. With a “fix first” verdict you get a list of changes to make yourselves or in an AI Rescue.

I reply personally

Message me on LinkedIn