Do people really choose lawyers with ChatGPT? What US surveys measure, and what they don't

What the new survey headlines claim

The headlines say consumers are moving from Google to ChatGPT to find a lawyer. A US consumer survey by iLawyer Marketing, released on 14 July 2026, asked people which online sources they would use to research a lawyer for an important legal issue, and reported a sharp year-on-year rise in AI answer engines.¹ iLawyer Marketing is a legal marketing agency that sells answer engine optimisation to law firms.

The survey is useful as a signal, and its own method section is candid about what it asked. The question was about the sources people would use. That is intent. It does not tell a firm how many clients chose it, or a competitor, because of an AI answer.

Our view: intent surveys are a fair reason to check what AI says about your firm. They are not a reason to rebuild a marketing budget around AI on their numbers alone.

What do these surveys actually measure?

They measure three different things, and headlines tend to merge them. The three are whether people have asked AI a legal question, whether they would use AI to research a lawyer, and whether AI influenced who they hired.

Three questions surveys ask, and what each tells a firm

Have used AI for a legal question

• Measured behaviour
• Often about the law, not about a lawyer
• May end with "talk to a lawyer"

Would use AI to research a lawyer

• Stated intent
• Easy to say yes to
• Rises with general AI use

AI influenced who was hired

• The question that matters to a firm
• Rarely asked
• Best measured in your own intake

Clio's 2025 Legal Trends Report, as reported by Canadian Lawyer in October 2025, combined consumers who had already asked AI a legal question with those who would only consider it.² It also looked at whether AI told the people who asked to contact a lawyer. Combining use and consideration makes a bigger headline; when you read figures like these, keep the two apart.

Asking AI what a tenancy law says is not the same as asking it to recommend a lawyer. A person who gets a clear answer may never need a firm at all. A person whose answer ends with "speak to a lawyer" is the one a firm wants to reach next.

Why intent is not hiring

Intent overstates behaviour for three reasons. People say yes to sources they might glance at. They count a quick check as research. And the decision to hire usually rests on other things, such as a referral, a consultation or fees.

Legal services also have long, high-stakes decisions. A person facing a divorce or a business dispute will often ask friends, read reviews, check a bar listing and call two or three firms. AI may appear early in that journey without deciding it. Surveys of "would use" cannot say where in the journey AI sat.

The best measure is in your own intake. Add "How did you first hear about us?" with an option for AI assistants, and ask new clients which sources they checked before calling. Within a few months you will have numbers for your own practice, which are worth more than any national survey.

Corporate clients are a clearer case

For firms serving businesses, the evidence is about the firm's own use of AI, not about search. The Thomson Reuters Institute's Future of Professionals 2026 report surveyed 1,816 professionals in 62 countries in March and April 2026. Among corporate respondents, 78% said receiving AI-enabled quality improvements from the firms they work with is very important or essential, and 6% said most or all of their providers deliver it.³

That gap points to a different job for a firm's website. In-house buyers want to know how a firm uses AI in its work, under what safeguards, and how that shows up in fees. A firm that publishes this plainly gives both buyers and AI answers something specific to repeat.

What should a firm publish?

Publish the facts a client needs to decide whether to call, in plain text, on pages that load without scripts. These are also the facts AI answers most often get wrong when a firm leaves them out.

• Practice areas, each on its own page, with the kinds of matters you take and do not take.
• Jurisdictions: the states where your lawyers are admitted and the courts where you appear.
• People: who handles which matters, with admissions and years of practice stated as facts.
• Fees: how you charge (hourly, flat, contingency), what a first consultation costs and what affects the total.
• Process: what happens after someone contacts you, and how long a first response takes.
• AI use: for business clients, how you use AI in your work and how client data is protected.

Fees are where firms hold back most, and where AI answers fill the gap with guesses. You do not need a price list. A clear explanation of how you charge, with typical ranges where your rules allow, gives an answer something accurate to say. Our guide to fixing what ChatGPT gets wrong about a company covers how to correct errors once you find them.

Why "best lawyer" wording backfires

Superlatives create two problems: one with the advertising rules, one with AI answers. Many US states model their lawyer advertising rules on the ABA Model Rules. The comment on Model Rule 7.1 warns that an unsubstantiated claim about a lawyer's services, or an unsubstantiated comparison with other lawyers, may be misleading if it is specific enough that a reasonable person would think it can be substantiated.⁴ State versions differ, so check your own bar's rules.

The AI problem is quieter. An answer engine looking for a reason to name your firm needs facts it can state: admitted in two states, handles commercial leases, offers a fixed-fee first consultation. "The best business lawyer in the state" gives it nothing to verify, so it either ignores the claim or repeats it without support, which is worse.

Our view: replace every superlative on your site with the fact that made you want to write it. If you are proud of trial experience, state the number of trials you have handled, if your rules allow. Specific, checkable facts are safer under the rules and more useful to an AI answer.

How to check what AI says about your firm

Run the questions your clients ask, in their words, and read the answers. Include practice-area questions with a state named, such as "Which firms handle commercial lease disputes in Texas?", and questions about your firm by name.

Record whether your firm is named, whether the facts are right and which sources the answer cites. Answers vary between runs, so ask each question more than once, on different days. Our buyer's spec for an AI visibility score explains how many runs a reliable number needs. Our audits run each prompt on three separate days across six engines, for that reason.