GEO 2026: how lawyers position themselves on AI engines

Your future clients no longer search Google the way they used to

In 2026, someone wondering how to challenge a dismissal no longer types "employment lawyer Paris" into Google. They put the question straight to ChatGPT, Google AI or Perplexity. And the AI answers, citing (or not citing) your firm.

Traffic from AI engines rose 527% in a year (Semrush, 2025). Google AI Overviews now reaches 2 billion users a month. ChatGPT counts 700 million weekly active users.

The critical point: 60% of Google searches no longer generate any click to a website. The AI answers directly on the results page. When an AI answer appears, only 8% of users click the classic results below it (Pew Research, 2025).

Why lawyers are on the front line

88% of the searches that trigger an AI answer are informational queries: questions of procedure, of cost, of timing, of risk. That is exactly what people with a legal problem are looking for.

A visitor who reaches your site through an AI answer is worth 4.4 times more than a classic visitor (Semrush, 2025). They are more qualified, more engaged, closer to getting in touch.

50% of the sources ChatGPT cites are professional services sites, not media outlets or forums. Law firm sites with structured content are squarely in the target.

Classic SEO vs GEO: the fundamental difference

Classic SEO places you in a list of 10 results. GEO (Generative Engine Optimization) places you in the answer itself. It is your content the AI rewrites and cites in front of the reader.

What the AIs look for before citing you: structured content, identifiable expertise, a direct answer and verifiable sources.

The "answer first" method

The AIs pull the first 2 or 3 sentences of a page to build their answer. If your page opens with a vague introduction along the lines of "French employment law is a complex field requiring sharp expertise", the AI moves on to the next site.

What works: opening every page with a direct answer in 2 or 3 sentences, citing the statutes involved, giving concrete figures (amounts, deadlines, percentages).

A concrete example: instead of "our firm supports you through your dismissal issues", write "a dismissal is considered unfair when it does not rest on a real and serious cause (article L1232-1 of the Labour Code). The employee has one year to bring a claim before the employment tribunal."

70% of readers only read the first third of an AI answer (Growth Memo, 2025). Be in that first third.

Tell the AIs who you are: entity signals

AI engines do not cite anonymous sites. They cite identifiable entities: a firm with a name, a bar, a city, a specialism, a qualified author.

The signals to put in place:

• Firm identity: full name, bar of registration, address, company number, repeated consistently across every page
• An identified author on every article: name, qualification, years of experience, fields of competence
• Verifiable local anchoring: mention the local courts, the specific procedures and the practical timings of your jurisdiction
• A visible update date: display "last updated" on every strategic page and refresh it every 3 to 6 months
• Verifiable, dated sources: cite the articles of the Labour Code, case law and decrees with links to Légifrance

The JSON-LD every law firm site should carry

Structured data helps Google and the AIs understand what your firm does, where it is, and why it is qualified. Google recommends it explicitly for producing rich results.

The schemas to implement first:

• LegalService: identifies your firm as a legal service, with name, area served, address, telephone
• Person: each lawyer with their name, title, specialisms (reinforces the attribution of expertise)
• FAQPage: on every topic page, structures your questions and answers so the AI can extract them directly
• Article + BreadcrumbList: on your blog articles, with author, publication date, modification date, category

The pages to create now

AI Overviews appear above all on long-tail queries: 68% have fewer than 100 searches a month and 80% have low ranking difficulty (Semrush, 2025). That is exactly the ground of precise legal questions.

Decision pages (high priority):

• "Is it worth challenging my dismissal?"
• "Do I need a lawyer for an uncontested divorce?"

Procedure pages:

• "How does an employment tribunal hearing work?"
• "Which documents for a settlement agreement?"

Cost and risk pages:

• "How much does an employment lawyer cost in 2026?"
• "What are the risks if I refuse a redundancy plan?"

Build topic clusters

The AIs assemble their answers from several sub-questions. A cluster architecture (a pillar page plus sub-pages) lets them move through your expertise like an encyclopaedia.

For example: a pillar page "Dismissal: