AI has moved in. Uninvited?

An experienced entrepreneur, fully aware of everything that goes on in his company, admitted to me, surprised and somewhat embarrassed, that he does not know exactly which AI tools have been deployed across his teams. We are working together on his brand positioning and were reviewing internal resources and processes.

That conversation echoed two episodes of ‘Changement d’époque en cours’ (an excellent French tech podcast hosted by Laurent Nicolas-Guennoc) that I listened to recently. In the first, Sébastien Le Roy, Partner at Serena Capital, unpacked the French AI Report, with one figure that shows just how fast things are moving: in 2025, Europe raised €12 billion for its later-stage AI companies, while the United States mobilised €141 billion! This wave of investment will sweep through every organisation, large and small. Many of them are not ready for it, because they do not yet know which AI tools are already in use.

AI has been settling into offices for months, adopted in everyday work, in translation tools, writing assistants, recruitment platforms and customer relationship systems, often at employees’ own initiative and without any real approval. That brings real risks of confidential data leaks or of personal data being processed.

In another episode, Debora Cohen, digital compliance expert at Converteo, former lawyer and a pioneer on digital issues, set out a major challenge: the urgent task for companies is to document their choices right now, including those made without their knowledge.

The entrepreneur I am working with is no exception. He is like a great many business leaders steering organisations in the midst of transformation without a complete map of what is in place.

Where to start? Debora Cohen recommends three concrete actions:

• draw up an inventory of the tools that incorporate AI,
• identify, for each one, who the provider is and who the deployer is within the company, then
• train teams so that every employee becomes a conscious link in the chain.

Knowing which AI systems are running in your organisation, on which data and with what access, has become a governance issue, just like cybersecurity.