AI audit: know where AI will help you, and where it will not
Your leadership wants to “do AI”, your teams already use assistants with no framework, and every vendor assures you its product solves everything. An audit puts things in order: what exists, what is possible with your data, what it would cost and where to start.
The problem
Without scoping, companies make two symmetrical mistakes. Some multiply demos that never reach production. Others wait, while their employees paste confidential data into consumer tools.
The right question is not “what can AI do?” but “which problems do we have, and for which is AI the cheapest answer?”. Sometimes the answer is a business rule, a better-designed form or an integration between two systems.
How we work
-
1
Listen to the business
Interviews with leadership and operational teams, observation of real work. We record repetitive tasks, bottlenecks and AI usage already in place, declared or not.
-
2
Look at data and systems
Where the data is, who can access it, what state it is in, what existing software allows you to integrate. Many AI projects are really data projects.
-
3
Sort the use cases
Each case is scored on four criteria: business value, feasibility with your data, risk (legal, reputational, operational) and effort. We write down the ones we recommend dropping, and why.
-
4
Deliver a roadmap that commits
Two or three priority projects with scope, budget, measurable success criteria and stop conditions. Plus a usage policy for employees and the points to settle with your DPO under the GDPR and the EU AI Act.
What makes these projects fail
An audit concluding you should do everything
A list of thirty use cases helps nobody decide. The value of an audit lies in the sorting, and in what it recommends not doing.
An audit run by whoever sells the solution
A vendor will conclude that its product is the answer. Our recommendation does not depend on any vendor, and we tell you when an off-the-shelf tool beats custom development.
Forgetting people
A technically successful project that teams work around is a failure. The audit also identifies who will own each project on the business side, and what will change in daily work.
What you receive
- Review of existing AI usage, declared or not
- Map of data and integration constraints
- Grid of use cases assessed, kept and dropped
- Prioritised roadmap with budgets and success criteria
- AI usage policy for employees
- Presentation to leadership
Frequently asked questions
How long does an AI audit take?
It depends on the size of the organisation and the number of business lines involved. We set duration and a fixed price after a first conversation, so the cost of the audit is known before it starts.
Do we need AI projects already under way?
No. The audit is useful to get started, to bring order to scattered initiatives, or to assess a project that is not moving forward.
Do you then carry out the recommended projects?
We can, but nothing obliges you to. The roadmap is written to be usable by your teams or by another provider. If you entrust us with the next step, we already know your context.
Is your situation close to this one?
Describe it in a few lines. We will tell you whether AI is the right answer — and we will also tell you when it is not.
Talk about your project