An AI project should start with decisions, not with tools.

AI audit and diagnosis

Take stock before investing further

You may already have experiments, tools in place or a project handed to a provider. The audit clarifies what actually works, what remains fragile and what must be decided before moving forward.

• use cases and value produced
• tools and solutions already in place
• affected processes
• visible and hidden costs
• dependencies on vendors or providers
• quality of data and knowledge used
• responsibilities and control mechanisms
• the system's ability to evolve

You get a clear reading of the current state, risks, priorities and next decisions.

Product and business framing

Turn a broad idea into an arbitrable project

We need to do something with AI is not enough to launch a project.

Framing turns ambition into a concrete proposal: for which users, in which situations, with which value and under which conditions.

• the business problem to solve
• users and processes involved
• priority use cases
• expected value
• AI's place in the experience
• the split between automation, assistance and human decision
• acceptable limits
• success criteria
• hypotheses to validate

You get a project that is understandable, shareable and precise enough to decide what comes next.

AI strategy and roadmap

Choose where to invest and in what order

AI initiatives multiply quickly: one tool per team, several prototypes in parallel, subscriptions piling up and no shared view.

I help you build a coherent trajectory from your business priorities, internal capabilities and maturity level.

• opportunity mapping
• comparison and prioritisation of use cases
• buy, integrate or build decisions
• identification of prerequisites
• organisation of experiments
• skills to mobilise
• ramp-up steps
• conditions for scaling

You get a realistic roadmap, with explicit choices and clearly ordered steps.

Solution design and architecture

Define how the system must actually work

An AI solution is not just a model choice or a chat interface.

You need to define which information it uses, what it can do, how it fits into existing work, when humans intervene and how its actions can be understood or controlled.

• user journeys and experience
• role of AI and role of humans
• information flows
• business rules
• levels of autonomy
• validations and exceptions
• integration with existing tools
• context and continuity management
• traceability of actions and decisions
• structuring principles for the architecture

You get a target vision clear enough to guide internal teams, partners and providers.