Advisory - François Bouet

Every legal-tech vendor now has an AI story. The demos are impressive, and they all sound the same. Underneath, some of these tools genuinely change how the work gets done. Many will quietly stall the first time they meet a real matter, a real document set, and a real confidentiality boundary.

I help legal teams tell the difference before the firm spends a year and a budget finding out. I am a software engineer by trade, not a reseller, so the read you get is on the mechanics: what a tool actually does, where it breaks, and whether it earns its place.

AI reality-check

A fixed-scope assessment of the AI in front of you. I map how your people actually work, evaluate the specific tools on the table against those workflows, and hand back a ranked verdict: what to adopt, what to pilot, what to ignore. Named tools, named owners, plain reasons. No hourly meter.

Grounded deployment over your own documents

Most legal AI answers from public caselaw or a model's general knowledge. The value is an assistant that answers from your work: your precedents, your closed matters, your templates, with citations you can check. I select, configure, and deploy existing tools so the answers are grounded in the firm's own corpus, not borrowed from the internet.

AI governance you can defend

The guardrails a regulated practice needs: what is safe to put into which tool, where privilege and confidentiality lines fall, what gets logged, and a usage policy partners will actually follow. The difference between using AI and being able to explain, later, exactly how you used it.

How I work

You work with me, not a pyramid of juniors. Engagements are fixed-scope, so you know the cost before we start. I handle the data-handling and confidentiality terms as part of the engagement. And I begin with a short discovery teardown of your actual situation before proposing anything, because the right answer depends on your matters, your stack, and your people, not a generic playbook.