Law Firm Appearance Rates in ChatGPT and Gemini
Law Firm Appearance Rates in ChatGPT and Gemini
1,450 prompt tests were run through the LegalVIS AI Prompt Checker across ChatGPT (724 tests) and Gemini (709 tests). Prompts simulate realistic legal service queries a prospective client would ask.
Appears rate is the proportion of individual prompt tests in which a firm was named. Recommended means the firm was actively suggested as a solution, not merely mentioned. Perplexity (13 tests) and Claude (4 tests) are excluded from headline comparisons due to low volume.
The 20 firms that dominate AI responses to legal queries.
The difference between being known and being retrievable.
The practical steps behind the appearance rate.
The buying journey these prompts are modelled on.
Independent research on how consumers search for and choose legal services.
Professional expectations when handling enquiries generated by any channel.
The AI visibility funnel: from search to recommendation
On any given AI legal search, roughly 1 in 4 audited firms appears, and fewer than 1 in 5 is recommended
236 firms tested · ChatGPT and Gemini · 1,450 prompt tests combined
On any given search, a 22.8% chance the firm is named. This is what a prospective client experiences.
Consistent with ChatGPT, roughly 1 in 4 searches surfaces the firm.
Appears rate per test: proportion of individual prompt tests where a firm was named, across 724 ChatGPT and 709 Gemini tests. Source: LegalVIS AI Prompt Checker, 2026.
The problem is not converting visibility into a recommendation, it is getting visible at all
122 firms that appeared in at least one AI response · ChatGPT and Gemini combined
Of the 122 firms that appeared in at least one AI response,
. Only 6 appeared without a recommendation.
The entire battle is fought at the visibility stage.
A firm that cannot get into the AI response has no chance of a recommendation.
Based on 122 firms that appeared in at least one AI response across ChatGPT and Gemini prompt tests. Source: LegalVIS AI Prompt Checker, 2026.
ChatGPT and Gemini are closely matched on visibility, 22.8% vs 23.3%.
The difference emerges at the recommendation stage: ChatGPT recommends at 22.2%, Gemini at 20.2%.
Rates calculated per individual prompt test. Source: LegalVIS AI Prompt Checker, 2026.
No firms at all sit in the 40–50% band. Firms appearing consistently are not marginally better, they are structurally different.
223 firms within the 335 deduplicated cohort with at least one prompt test run. Source: LegalVIS AI Prompt Checker, 2026.
Because 95.1% of firms that appear are then recommended, work spent on making the firm sound more persuasive has almost no effect on AI outcomes. Work spent on being retrievable has all of it.
Retrieval depends on structured signals:
A prompt test on your own firm, using the queries your clients actually ask, is the fastest way to see which side of the bimodal split you sit on.
Further reading from LegalVIS
Sources and further references
About the researcher
“The shift to AI search is already changing how people discover and shortlist professional services. Ambitious law firms should be thinking about how they become part of those recommendations now, not waiting until AI visibility becomes standard practice.”
Claire Mullaney, Founder, LegalVIS
Claire founded LegalVIS, the UK’s first AI visibility platform built specifically for the legal sector. She works directly with law firms across the UK and internationally on structured AI visibility, and works closely with LAWorld, the international legal network, on AI visibility strategy.