Bredin Prat owns ChatGPT. Every other Paris firm is playing in the second division.

On ChatGPT, Gemini and Perplexity, one French firm dominates business law: Bredin Prat, with an AI visibility score of 100 out of 100.

Ask ChatGPT, Gemini or Perplexity to name the best business law firm in France. You'll get more or less the same answer every time, in the same order. Bredin Prat. Then Gide. Well behind them, Darrois Villey. And further back still, a British firm, followed by a legaltech startup.

This isn't an opinion. It's the raw output of the KAIS 2026 barometer, which ran thirty legal queries — fifteen framed from the point of view of a client looking for a lawyer, fifteen from the point of view of a market player such as an investor or general counsel — through the three most-used LLMs in France. 1,870 citations analysed. A single ranking, the KAIS Visibility Score, rebased to 100 for the sector leader.

Historic authority no longer guarantees algorithmic authority

The finding cuts against a widespread assumption. On paper, Linklaters and Clifford Chance dwarf most French firms in revenue, in deals closed, in headcount. In Chambers and the Legal 500, they reign. Among their peers, they're the benchmark. To French-language LLMs, they barely exist. Their editorial output stays Anglo-Saxon, their communications aimed at professional rankings, their Schema.org markup limited to a corporate profile. The models don't know where to find them in French.

Bredin Prat, by contrast, dominates for a reason that has nothing to do with its reputation as a dealmaking powerhouse. The firm gets cited consistently across both layers of the test. When an individual asks which firm is the most respected in France, Bredin Prat comes up. When an investor searches for the best business law firms in Europe, Bredin Prat comes up again. The study calls this multi-layer consistency, and it's where the score is built.

AI visibility isn't additive across layers. It's multiplicative. A firm spotted only by the general public, like Captain Contrat, or only by market players, tops out somewhere between 12 and 20 points. To break 50, you have to hold both fronts at once.

What LLMs actually look at

Three factors show up in every sector analysis in the barometer. None of them reduces to classic SEO.

1. Entity resolution

The models don't read a website in isolation. They cross-reference the firm's profile, partners' LinkedIn pages, bar directories, press mentions, professional rankings. If a partner is presented as an M&A specialist on the firm's site but posts about employment law on LinkedIn, the model detects a mismatch and drops the source as a precaution. Semantic consistency across all these points matters more than sheer volume published.

2. Precision of vocabulary

"Business lawyer" has become an empty term for LLMs — saturated, with no discriminating value. The firms that have climbed over the past eighteen months write differently: "cross-border intellectual property litigation in the SaaS software sector," "post-acquisition disputes under German law," "consumer class actions under European law." The closer the vocabulary sits to a real use case, the more readily the model finds its answer there.

3. The authority graph

LLMs extend trust to sources cited by other recognised sources. An op-ed in Les Échos, an analysis in Dalloz, a contribution to a leading legal review carry more weight than a post on the firm's own blog. It's the difference between claiming to be an expert and being named as one by an external node of authority.

robots.txt, the profession's blind spot

One technical detail deserves attention. Many firms blocked OpenAI's crawler, GPTBot, between 2023 and 2024, fearing their analyses would be scraped and served up without any return. The instinct is understandable. The result, two years on, is that they're no longer indexed at all by the sources ChatGPT consults in real time through SearchGPT. A misconfigured robots.txt now costs more than the intellectual property it was meant to protect.

Allowing the crawl without tagging your pages with the right microdata produces the opposite but equally costly effect. The model sees the content but has no way to interpret it. Schema.org's LegalService markup, paired with Attorney and FAQPage, tells the machine explicitly: this is an article written by a lawyer registered with the Paris bar, specialising in business law, analysing a given ruling of the Court of Cassation. The code stays invisible to the human reader. To the model, it's the difference between a page that's legible and a page that's indexable.

What a firm ranked between 10 and 30 can do

The most useful lesson in the barometer concerns neither Bredin Prat nor Gide, whose positions are locked in for the next two to three years. It concerns every firm ranked between 10 and 30 — real expertise, AI visibility close to zero.

For these firms, producing more generalist legal content is a dead end. The ground is saturated. The lever that pays is the precise vertical. Cybersecurity and data protection. Renewable energy. International succession. Healthcare M&A. AI Act compliance. A firm that publishes two or three times a week on a single one of these niches, with the right technical markup and a consistent presence on LinkedIn and in the specialist press, can move from a score of 5 to a score of 40 on its vertical in nine months. LLMs cite what they find, not what's prestigious.

The first results land within thirty to sixty days, the time it takes SearchGPT and the Perplexity index to refresh their reading of the sources. Owning a niche takes three to six months to build. For a focused firm, showing up in forty to fifty answers a week on high-intent commercial queries represents a flow of qualified prospects no brochure site will ever generate.

Two or three years from now, the Chambers and Legal 500 rankings will keep coming out. The question is how many clients will still read them, and how many will simply have asked a model to recommend them a lawyer.