AI governance and GDPR-compliant use
Bordeaux and Nouvelle-Aquitaine, remote across France
Your teams already use AI tools, with or without your approval. The question is no longer whether to allow them, but which ones, for whom, with which data. I step in to put AI governance in place that holds: a short usage charter, a data classification that says what may be submitted to a tool and what never leaves, a register of uses, and teams who know the rules. All of it built with your DPO, not around them.
Book a free callThe problem I solve
The scenario I meet most often: a blanket ban, never enforced, and staff pasting a client contract or an extract of the HR database into a consumer assistant to save an hour. Nobody knows, so nobody can prevent it or document it. The day a client, an auditor or the regulator asks the question, there is no rule, no register and no trace of any awareness session. The problem is not the tool, it is the absence of a framework. I build it with you: data categories everyone understands, uses that are authorised by name, and a register kept up to date.
How I work
Inventory of current uses
Before writing a single rule, I look at what actually happens: which tools are used, by whom, for which tasks, with which data. Short interviews per team are enough. That picture becomes the first version of the register of uses.
Data classification
I define with you three levels anyone can read: what may be submitted to a tool as is, what must be anonymised first, what never leaves. Each level comes with examples drawn from your own documents.
Charter and register
The charter runs to a few pages and says who may use what, how to check an output, and whom to alert. The register of uses lists each authorised case, its tool, its data and its owner. Both are reviewed by the DPO.
Team awareness
A rule nobody knows protects nobody. I run short sessions per team, on their real cases, with examples of what passes and what does not. Then I plan the reminders and the updates.
Real example
Where this comes from: at La Banque Postale, since February 2025, I have been steering account opening journeys under permanent KYC and GDPR constraints, with a budget of around 1 million euros, 2 external partnerships integrated through APIs and awareness sessions for the teams on data governance. These steering results are not a promise, they say where I speak from.
Read the case studyWho it is for
- IT departments that must authorise AI tools without exposing company data
- DPOs looking for an operational relay to get the rules applied in the teams
- Business units whose teams already use AI with no written framework
- Banks and public sector bodies subject to audits on how they use data
What you get
- A short AI usage charter, reviewed with your DPO
- A three-level data classification, with examples from your own documents
- A register of AI uses, kept up to date and open to consultation
- Teams made aware on their own cases, with a reusable support deck
- A review plan: who updates what, and how often
Frequently asked questions
Is this legal advice?+
No. I produce neither legal analysis nor compliance opinions: that framing stays with the DPO or the lawyer. My work starts afterwards: turning their requirements into rules the teams understand, into data categories usable day to day, and into a register that really exists. I work with them, not in their place.
What hands-on experience do you have with GDPR?+
Since February 2025 I have been steering the digital onboarding journeys of a major bank, under permanent KYC and GDPR constraints, including awareness sessions for the teams on data governance. Before that, I steered the migration of 500 million documents in GDPR compliance. I have also taught a course on generative AI and GDPR at an IT school.
Which AI tools can be authorised?+
It depends on your data and your contracts, not on a universal list. A consumer assistant, an enterprise version with a commitment not to reuse your data, and a model hosted on your own infrastructure do not offer the same guarantees. For each tool under consideration I document where the data goes and who can access it, and the decision is made on that basis.
How much does it cost?+
On a quote basis, depending on the number of teams to cover and whether you have an in-house DPO. A typical assignment includes the inventory, the drafting of the charter and the classification, then the awareness sessions. The first call is free.
The programmes
This service is not legal advice. Legal framing and compliance opinions remain the responsibility of your DPO or your lawyer; I turn them into operational rules within the teams.
An AI usage framework to set
Book a free first call. You describe your tools and your data, I tell you where to start and what belongs to your DPO.