KAIS Barometer · AI Visibility · 2026 EditionÉd. 1 · 3 engines

Law firms in Paris & Lyon, as the AIs see them

An observational study of law-firm visibility in Paris and Lyon across Gemini, Meta AI and Perplexity. The results describe citations observed across 80 legal queries at the collection date; they do not measure legal quality, professional competence or case outcomes.

01 — KEY FINDINGS

Six observations from the legal-services panel

Six patterns emerge from 80 real-world legal queries across three engines. The panel contains 240 question × engine cells, each measured three times; 640 citations were recorded in the consolidated dataset used for the ranking.

02 — RANKING

The AI Visibility ranking (KVS)

The KAIS Visibility Score is a relative visibility index within this panel, weighted by citation rank and multi-engine coverage and then normalised to 100 per city. It is not a probability of recommendation, market share or assessment of legal quality.

■ Gold bar = the city's absolute leader (KVS 100) KVS normalized per city and practice area — panel: Gemini · Meta AI · Perplexity
03 — THE 3-ENGINE RULE

One engine isn't enough: three behaviors, three audiences

The three engines name neither the same number of firms nor the same ones. A firm can be perfectly visible on Gemini and absent from Meta AI for the same question. Optimizing for a single engine caps your visibility.

The cost wall:

04 — BY PRACTICE AREA

Who is most visible by practice area

The hierarchy changes radically from one practice area to the next — and from one city to the other. The Reference Rate is the share of answers that name at least one firm in the area.

05 — PER-QUERY ANALYSIS

High-intent queries, fragmented areas and queries with no firm cited

This view separates high-intent queries where one firm repeatedly leads, queries where leadership varies by engine, and queries where no firm is cited. It is designed to prioritise editorial and authority experiments; it does not describe reserved positions or guaranteed opportunities.

Fine per-query read: ranking of cited firms and their position on the Gemini + Meta AI text sub-panel (Perplexity remains included in the aggregate KVS). "Recurrent leadership" means a firm repeatedly appeared at the top of the observed sub-panel. "Fragmented" means leadership differs by engine. These labels are descriptive and may change over time.

Methodology

Limitations. This is an observational study of engine outputs at a specific point in time. It does not establish that a particular page, source or action causes a citation and is neither a ranking of legal quality nor a recommendation of a law firm.

KaiZen AI Strategy

Is your firm cited on your queries?

KAIS works with law firms that need to measure and strengthen their visibility in AI search, using a market-appropriate engine panel, archived measurements and a framework compatible with professional-conduct requirements. No position or citation is guaranteed.

View KAIS diagnostics →
Ed. 1 · 3 engines Legal KVS Barometer — Law firms, Paris & Lyon 2026. Data: 80 queries × 3 AI engines (Gemini · Meta AI · Perplexity), recorded on 30 June 2026; 240 question × engine cells, three runs per cell. KVS is a relative visibility index within this panel, not an assessment of legal quality. Firm names are shown as they appeared in the AI answers. Source: Legal AI Barometer 2026, KaiZen AI Strategy.