Situations
Where to start with AI in a small company
You know AI could help somewhere, but not which process to point it at first. It is the most common situation, and also the one that resolves fastest.
What we observe
- Scattered ChatGPT experiments across the team, nothing running continuously
- Leadership asking "where are we on AI" with no specific target
- Several use case ideas, none of them costed
- Nobody can say how much time the target task costs today
What usually blocks
The blocker is almost never technical. It comes from nobody having measured the process before discussing it, so every idea looks equally good and none gets picked.
A useful first case has three properties: steady volume, verifiable output, and someone on the team who suffers from the task enough to describe it precisely. Miss one of the three and the project stalls, however well built.
What we do
- 01We list the real frictions with the people living them, not with the leadership team
- 02We measure for two weeks: volume handled, time per unit, rework rate
- 03We rank cases by gain over effort and keep exactly one
- 04We scope that case with its guardrails and a checkpoint at three months
The method in full
The reference guide on the subject, with its definition, its sources and its guardrails.
Scope your first AI agent
20 minutes to review your tools, data and the first useful case. No jargon, no commitment.