An AI implementation plan in four steps, from diagnosis to adoption
A method that starts from your real processes, ranks use cases by value and risk, then puts them into production with the teams who will use them.
The four steps
Each step ends with a written deliverable, validated with you before moving to the next.
- 01
Scoping and diagnosis
We start from your processes as they run: where time goes, where the data lives, where decisions are made and by whom.
Deliverable
A map of processes and available data
- 02
Use case prioritisation
Each use case is scored on three criteria: expected value, technical feasibility and risk, including sensitive data, compliance and dependencies.
Deliverable
A prioritisation grid validated by management
- 03
Roadmap and governance
A phased timeline, named owners and usage rules: what is allowed, with which data, and who validates the outputs.
Deliverable
A roadmap and an AI usage charter
- 04
Rollout and capability building
The first use cases go into production on real files, with team training and follow-up after launch.
Deliverable
The first deployed use cases and a training plan
How progress is tracked
Quantified targets are set with you, from a measured starting point. We do not write them before the diagnostic.
- Time spent on targeted tasks, before and after rollout
- Adoption rate by team and by process
- Output quality, with the human review set out in the charter
- Compliance with the usage charter and incidents reported
The starting point is a diagnostic
It shows your real situation before anything is decided. The plan is then built with you.
Book a call