Deutsch
Français
Español
Português (Brasil)
ItalianoWhich criminal defense brands
AI recommends in the US
Brands ranked by mention rate on ChatGPT, Gemini and Google AI Mode
- Brands
- 85
- Questions
- 400
- Answers
- 1,200
- AI engines
- 3
01 · Brand leader
Top 10 brands by mention rate
No single brand runs away with the category.
Several brands compete for the top places.
| # | Brand | Mention rate Share of questions where ChatGPT names the brand | Share of voice The brand’s share of all brand mentions in the engine’s answers | 95% range The interval the true mention rate most likely falls in | In top 20 of The brand’s place inside each engine’s own top 20 |
|---|---|---|---|---|---|
| 11–6 | Williams & Connolly | 12.8% | 7.1% | 9.5–16.0 | ChatGPT1Gemini1Google AI Mode1 |
| 21–9 | Debevoise & Plimpton | 10.5% | 5.8% | 7.5–13.5 | ChatGPT2Gemini12Google AI Mode5 |
| 31–12 | Covington & Burling | 8.0% | 3.4% | 5.7–11.1 | ChatGPT3Gemini4Google AI Mode4 |
| 41–12 | Latham & Watkins | 7.8% | 4.1% | 5.5–10.8 | ChatGPT4Gemini2Google AI Mode3 |
| 51–12 | Eisner Gorin | 7.5% | 11.1% | 5.3–10.5 | ChatGPT5Gemini8Google AI Mode11 |
| 61–12 | Gibson Dunn | 7.2% | 4.0% | 5.1–10.2 | ChatGPT6Gemini5Google AI Mode10 |
| 72–15 | Spodek Law Group | 6.0% | 10.1% | 4.1–8.8 | ChatGPT7Gemini17Google AI Mode19 |
| 82–15 | Miller & Chevalier | 5.8% | 2.8% | 3.9–8.5 | ChatGPT8 |
| 92–17 | Steptoe | 5.0% | 3.1% | 3.3–7.6 | ChatGPT9 |
| 103–20 | Quinn Emanuel | 4.3% | 1.7% | 2.7–6.7 | ChatGPT10Gemini11Google AI Mode6 |
| 113–20 | Morvillo Abramowitz Grand Iason & Anello | 3.8% | 1.5% | 2.3–6.1 | ChatGPT11Gemini9Google AI Mode8 |
| 123–20 | Venable | 3.5% | 2.0% | 2.1–5.8 | ChatGPT12 |
| 137–20 | Sullivan & Cromwell | 2.8% | 1.5% | 1.5–4.9 | ChatGPT13 |
| 147–20 | Broden & Mickelsen | 2.5% | 3.6% | 1.4–4.5 | ChatGPT14 |
| 157–20 | Davis Polk | 2.5% | 1.0% | 1.4–4.5 | ChatGPT15Gemini18Google AI Mode7 |
| 169–20 | Paul Hastings | 1.8% | 0.8% | 0.9–3.6 | ChatGPT16Gemini10 |
| 179–20 | Paul Weiss | 1.8% | 0.7% | 0.9–3.6 | ChatGPT17 |
| 1810–20 | Brafman & Associates | 1.5% | 1.7% | 0.7–3.2 | ChatGPT18 |
| 1910–20 | Nixon Peabody | 1.5% | 1.2% | 0.7–3.2 | ChatGPT19 |
| 2010–20 | Bracewell | 1.3% | 0.7% | 0.5–2.9 | ChatGPT20 |
02 · Engine consensus
Where each brand ranks
inside every engine’s top 10
There is no single AI top 10.
Only 4 of 17 brands rank in the top 10 of all 3 engines.
| Brand | ChatGPT | Gemini | Google AI Mode | Perplexity | In top 10 of |
|---|---|---|---|---|---|
| Williams & Connolly | 1 | 1 | 1 | n/a | ChatGPTGeminiGoogle AI Mode |
| Latham & Watkins | 4 | 2 | 3 | n/a | ChatGPTGeminiGoogle AI Mode |
| Covington & Burling | 3 | 4 | 4 | n/a | ChatGPTGeminiGoogle AI Mode |
| Gibson Dunn | 6 | 5 | 10 | n/a | ChatGPTGeminiGoogle AI Mode |
| Kirkland & Ellis | – | 3 | 2 | n/a | GeminiGoogle AI Mode |
| Debevoise & Plimpton | 2 | – | 5 | n/a | ChatGPTGoogle AI Mode |
| Eisner Gorin | 5 | 8 | – | n/a | ChatGPTGemini |
| Quinn Emanuel | 10 | – | 6 | n/a | ChatGPTGoogle AI Mode |
| Morvillo Abramowitz Grand Iason & Anello | – | 9 | 8 | n/a | GeminiGoogle AI Mode |
| Skadden | – | 6 | – | n/a | Gemini |
| Spodek Law Group | 7 | – | – | n/a | ChatGPT |
| Kobre & Kim | – | 7 | – | n/a | Gemini |
| Davis Polk | – | – | 7 | n/a | Google AI Mode |
| Miller & Chevalier | 8 | – | – | n/a | ChatGPT |
| Steptoe | 9 | – | – | n/a | ChatGPT |
| King & Spalding | – | – | 9 | n/a | Google AI Mode |
| Paul Hastings | – | 10 | – | n/a | Gemini |
Rank inside each engine’s own top 10 · a dash means the brand is not in that engine’s top 10 · n/a — the engine gave too few answers here to publish a ranking.
The rest depend on which engine the buyer opens — so a rank from a single engine is not a market position.
Source: Findrix's research. Where each brand ranks inside every engine's top 10, criminal defense, United States.
03 · Most cited domains
Which sites each engine cites, and
how much of its answers they carry
| Domain | ChatGPT | Gemini | Google AI Mode | Perplexity |
|---|---|---|---|---|
| chambers.com | 4.7% | 1.3% | 3.4% | n/a |
| vault.com | 0.3% | 4.6% | 3.1% | n/a |
| wc.com | 0.6% | 4.4% | 0.9% | n/a |
| lw.com | 0.4% | 3.2% | 0.5% | n/a |
| bestlawfirms.com | 0.8% | 0.6% | 2.5% | n/a |
| americanbar.org | 2.4% | – | 1× | n/a |
| federal-criminal-lawyer.com | 0.1% | 2.2% | 0.6% | n/a |
| lawyerlegion.com | 0.1% | 2.2% | 0.3% | n/a |
| cdllegal.com | 0.2% | 0.3% | 2.1% | n/a |
| gibsondunn.com | 0.4% | 2.1% | 0.3% | n/a |
| ca.gov | 1.8% | – | 0.2% | n/a |
| justice.gov | 1.7% | 1× | 0.2% | n/a |
04 · The questions we asked
400 buyer questions, each
asked once per engine
General requests and requests with a specific use case, audience, selection criterion or problem.
| Question type | Questions |
|---|---|
| General questions | 6 |
| Use case | 209 |
| Audience | 20 |
| Selection criteria | 120 |
| Problem to solve | 45 |
By question type
Examples · General questions
“what are the best criminal defense firms in the US?”
“which criminal defense firms are best value for money in the US?”
Examples · Use case
“which criminal defense firms are best value for money for a formal Department of Justice target letter in the US?”
“which criminal defense firms specialise in an internal audit discovering suspected embezzlement or securities fraud in the US?”
Examples · Audience
“which criminal defense firms are best for families trying to bail out loved ones facing serious felony charges in the US?”
“which criminal defense firms suit first-time DUI offenders facing license suspension best in the US?”
Examples · Selection criteria
“how do criminal defense firms compare on defense teams led by former Assistant US Attorneys in the US?”
“what are the top-rated criminal defense firms for offering flexible retainer installment plans for cash-strapped families in the US?”
Examples · Problem to solve
“cannot locate inmate in county jail booking system. which criminal defense firms can help in the US?”
“FBI agents question me without a lawyer present. which criminal defense firms can help in the US?”
Source: Findrix research. 400 buyer questions by question type, criminal defense, United States.
Neighbouring markets · Legal Services
United States · Australia
Family Law
ChatGPT leader Aronson Mayefsky & Sloan · 11.5% · 39 brands measured
Open this marketUnited States · Australia
Personal Injury & Compensation Law
ChatGPT leader Beasley Allen · 13.0% · 55 brands measured
Open this marketUnited States · Australia
Residential Conveyancing Services
ChatGPT leader First American Title · 48.5% · 25 brands measured
Open this market
Frequently asked questions
What does this Criminal Defense AI Search report measure?
Which criminal defense brands ChatGPT, Gemini and Google AI Mode name when buyers in the United States ask them for advice. We asked 400 buyer questions, each once to each engine, and counted how often every brand appears and which websites the engines cite. Measured Oct 2026.
What does “mention rate” mean?
The share of questions where an engine names the brand in its answer. A 10% mention rate means the brand came up in about one question in ten. Every rate carries a 95% range: the interval the true value most likely falls in.
How are the top Criminal Defense brands ranked?
By mention rate, separately for each engine — pick the engine above the table. When two brands’ 95% ranges overlap, their order is not really different — tick “Only distinguishable places” to see the ranks the data can actually separate.
Why does a brand rank differently across AI engines?
Each engine retrieves from different sources and weighs them differently, so the brands it knows best differ too. Only 4 of 17 brands make the top 10 of every engine here; the rest depend on which engine the buyer opens.
What counts as a citation?
A link or source an engine shows alongside its answer. We record every cited domain, whether it appears inline or in the source list.
Why can AI mention a brand without citing its website?
Engines learn brand names from reviews, expert sites, retailers and forums, then cite those sources rather than the brand’s own pages. A brand can be named often while its own site is rarely among the sources.
What does “share of citations” mean?
Of all the citations one engine gave across our questions, the share that went to a given domain. The shares of one engine add up to 100%.
Will I get the same answers if I ask these questions myself?
Not word for word. AI answers vary between runs, accounts and locations, which is why we ask every question once to each engine and report a 95% range instead of a single answer.