The Diagnostic
The starting point of everything. A documented verdict on your AI use cases — not a demo, not a promise.
Typically for an SME of 20 to 200 employees
Workflow observation
We observe your real processes, your exceptions and your systems — from documents and interviews, not from a questionnaire.
Proprietary scoring
Every use case is scored on observable sector criteria — a transparent method, no black-box score.
Verdict by use case
Launch first, framed pilot, or not yet — with the justification for each.
Client follow-ups
Strong impact, data ready.
LaunchAutomated quotes
Net time saved.
LaunchLevel 1 support
To frame before scaling.
PilotStock forecasting
Insufficient data.
Not yetThe report
A decision dossier, not a slide deck. Mapping, scoring, verdict, plan — in a document your committee can read and challenge.
To launch
2 priority cases
LaunchPilot
2 cases to frame
PilotThe 90-day plan
The first workstream, costed, with its owner and the proof threshold that decides what comes next.
How a verdict reads.
Atelier Zellige is a fictional client, used here to illustrate the shape of a diagnostic. Of eight use cases assessed, two deserve to be launched immediately, two move to a framed pilot, and four wait — for lack of data or sufficient impact.
The verdict is not "deploy AI everywhere." It shows where to start, why, and when to move on.
After the verdict, what follows.
The diagnostic opens four possible paths — you choose according to your priorities.
Engagement
An execution framework month after month, to hold course after the diagnostic.
Learn more → P3 — ExecutionManaged agents
Automations built, operated and supervised by the firm.
Learn more → P4 — TrainingTraining
Your teams upskilled on the tools you keep.
Learn more → P5 — LeadershipAI leadership
Fractional steering at leadership level, to frame the trajectory.
Learn more →