AI audit: measure your organisation's maturity and decide where to invest

Three kinds of firms offer it: general consultancies, which fold it into a broader transformation; IT services firms, often tied to an integration offer; and AI-specialist firms whose consultant masters the tools they recommend. Colombani.ai belongs to the latter: the audit is led by Ulysse Trin, who holds the Claude Certified Architect Professional certification issued by Anthropic and nine practitioner certifications issued by OpenAI, with no vendor to promote.

Colombani.ai uses five levels, the same as the online assessment: discovery (individual use without a framework), experimentation (volunteers and chosen tools), structuring (a shared framework to write and enforce), integration (AI in processes, its value tracked) and autonomy (the organisation designs and runs its own tools). The audit places the organisation on this scale and sets out what it takes to reach the next level.

An AI audit assesses an organisation's maturity, that is, its ability to use artificial intelligence in its work. It connects business objectives with available data, processes, skills, culture, and infrastructure. The result is a gap assessment and a roadmap for deciding what to test, prepare, or postpone.

One week for the Flash Diagnostic, two weeks for the full audit. The Flash Diagnostic covers key processes; the full audit covers all five areas and the departments selected during scoping. The interview schedule is set during scoping, depending on the availability of the people involved.

No. The audit starts with interviews and an analysis of the current state. Even without a centralized database, we identify use cases and missing prerequisites.

The full audit provides a maturity mapping across 5 areas, 5 to 10 use cases identified during the audit and prioritised, with their expected return, a gap analysis, and a costed roadmap with quick wins and strategic initiatives. The Flash Diagnostic provides three to five use cases to test, ranked by expected gain and effort, a short memo and a one-hour debrief.

Identify the teams involved, the tools they use, and two or three tasks you want to improve. Process examples, a list of data sources, and internal AI usage rules help scope the interviews. Relevant documents and sharing arrangements are agreed before collection; no sensitive case files are needed for the first discussion.

In a social care organisation, the assessment should distinguish support functions from care delivery. Searching internal procedures, preparing materials, and administrative reporting are possibilities to assess against actual needs. Each requires clarity on the data involved, authorised access, and expected human review. Uses affecting care decisions need specific scoping with service leads and those responsible for data protection.

AI maturity is assessed across five axes: data, processes, skills, culture, and infrastructure. Each axis is scored on observed facts, such as work examples, actual data access, and rules in force, rather than stated intentions. The assessment also records uses already in place, including those that go through no approved tool. A score only matters if it leads to a decision: what can be tested now, what needs preparing, and what should wait.

A digital maturity audit measures how far processes and tools are digitised; a data maturity audit examines data governance, quality, and architecture. An AI maturity audit builds on both but asks a different question: which tasks AI can take on today, with what data, under what human oversight, and with what risks. Where a digital or data audit already exists, its findings are reused rather than redone.

The Flash Diagnostic, two days of work over one week, costs €3,000 excl. VAT; the amount is deducted in full from the next consulting engagement signed within sixty days of the debrief, never from a training course. The full audit, carried out in two weeks, is quoted according to the size of the organisation and the number of departments interviewed. The first 30-minute call is free.

The online AI assessment is a free first reading: about a dozen questions, three minutes, and a level shown at once, at colombani.ai/en/ai-assessment/. The AI audit is the in-depth study: it relies on interviews and real work examples, and ends with a costed roadmap. Between the two, the Flash Diagnostic is a one-week study of key processes and first use cases; the full audit covers the five areas and the departments selected during scoping, in two weeks.