By industry · Legal

How Law Firms Show Up in ChatGPT and AI Search

Someone just got served, or rear-ended, or an offer on a house fell through. They ask an assistant which lawyer to call. Here is what they ask, why firm websites fail the machine that answers, and what to fix in thirty days without crossing an advertising rule.

Jason Colapietro · Suede AI · 21 September 2026

A law firm shows up in ChatGPT, Gemini, Perplexity and Google's AI Overviews when the engine can fetch the firm's pages, identify one firm in one place with named attorneys, quote a plain sentence about a specific practice area, and corroborate it on the bar's member record, Google Business Profile and Avvo. Law firm websites are, as a class, the worst-read pages we see: the copy is written to satisfy an advertising rule and a partner's taste rather than a question, so it says "results-driven advocacy" where the engine wanted "we handle contested divorce in Palm Beach County". Nothing here requires a claim you could not make in a bar-compliant ad. It requires saying the plain thing.

What a prospective client actually asks an AI

Legal questions to an assistant are specific about the problem and vague about the law, the opposite of how firm websites are written. Test six of these with your county in place of the placeholder and keep the screenshots.

  • "I was rear-ended on I-95 in Palm Beach County, do I need a lawyer and who is good?"
  • "Best divorce lawyer in West Palm Beach for a high-asset case"
  • "Immigration lawyer in [city] that speaks Spanish or Creole"
  • "How much does a DUI lawyer cost in Florida and who handles Palm Beach County cases?"
  • "Estate planning attorney near Jupiter for a trust, what does it cost?"
  • "Is [firm name] a real law firm? Reviews and bar record"
  • "Landlord will not return my deposit in Florida, what lawyer handles that?"
  • "Business lawyer in Boca Raton for an LLC operating agreement"
  • "Personal injury lawyer that works on contingency near Delray Beach"
  • "Who handles probate in Palm Beach County and how long does it take?"

Why the typical law firm website fails the machine reader

Entity clarity

"Smith & Jones", "Smith Jones Law Group, P.A.", "The Smith Law Firm" and a legacy domain from a former partnership all describe one firm, and the bar record carries the attorneys' names rather than any of them. An engine resolving "who is this" gives up quietly. One firm name, the legal name beside it, one address, one phone, and each attorney as a named person linked to their bar profile, on every surface the firm controls.

Crawler access

Firm sites built by legal-marketing vendors often sit behind aggressive bot protection, and the vendor's platform may block AI crawlers by policy. If OAI-SearchBot or PerplexityBot is refused at the door, the firm is not a candidate for the answer. The ten-second check at optimize.suedeai.ai shows the per-bot result. Access is only the gate; an allowed crawler proves no citation.

Money-page readability

The practice-area page opens with the firm's commitment to excellence. The plain version: "We represent people injured in car accidents in Palm Beach, Martin and St. Lucie counties, on contingency, with no fee unless we recover." When the fee statement is accurate that is a compliant sentence in most jurisdictions, and it is one an engine can quote. Each practice area needs one, a short FAQ in the client's words, and a visible updated date.

Evidence

The receipts a client and an engine both trust: bar number and admission year, courts admitted to, a real biography with a photo, board certification if it exists, and case results stated only as your bar's rules allow. Superlatives are the trap. Many bar rules restrict "best" and "expert"; plain facts with a source are safer and read better to a machine.

Reviews and citations

For "which lawyer" questions engines lean on Google reviews, Avvo, the state bar's member lookup, Justia and Martindale. A firm with 15 reviews and no Avvo profile is cited after one with 200 and a complete one. Asking a satisfied client for a review, offering nothing for it and within your bar's rules, is the highest-leverage slow item on this list.

Schema

Use LegalService (or Attorney for a solo) with name, address, telephone, areaServed as a list of counties, each attorney as a Person node, and a FAQPage block matching the questions visibly on the page. Structured data does not earn the citation; it stops the engine guessing which firm and where.

A 30-day fix sequence for a law firm

  1. Days 1 to 3: capture the baseline. Eight of the questions above with your county, on ChatGPT, Gemini, Perplexity and Google with AI Overviews. Screenshot and date. Record every firm named and every source cited.
  2. Days 4 to 7: open the gate. Read robots.txt and the bot-protection settings. Unblock the search crawlers and user-request fetchers. Confirm every practice-area page is in the sitemap with its copy in the initial HTML.
  3. Days 8 to 14: settle the entity. One firm name, legal name beside it, one address and phone. Every attorney with a bio page linked to their bar profile. Same strings on Google Business Profile, Avvo, Justia and LinkedIn. Add the LegalService schema block.
  4. Days 15 to 21: rewrite the practice-area pages. One plain first paragraph each: who you represent, where, on what fee basis, what happens first. A short FAQ and a visible updated date. Have the compliance reviewer read it; plain facts rarely trouble them.
  5. Days 22 to 28: build the record. Ten review requests within the rules. Avvo and Justia profiles completed. Any listing naming a departed attorney corrected.
  6. Days 29 to 30: measure again. Same questions, same engines, same method. Compare the sets and keep both.

The firms that get named are rarely the largest. They are the ones whose pages answer the question the client asked, in the client's words, with a bar record behind them.

Jason Colapietro is the founder of Suede AI in West Palm Beach, Florida, and runs the practice at seo.suedeai.ai. Every AI answer described here is a point-in-time capture of a third-party engine, and engine output changes without notice.

Questions

Why does ChatGPT recommend other law firms and not ours?

Because the engine retrieved sources that named them and could not retrieve or read sources that name you. For law firms the usual causes are bot protection or a vendor platform blocking AI search crawlers, a firm name that differs across the website, the bar record and Google Business Profile, practice-area pages written as slogans, and a thinner review and directory record.

Can a law firm do this within bar advertising rules?

Yes. Everything here is plain, verifiable fact: who you represent, where, on what fee basis, which courts you are admitted to, your bar number. Superlatives and unqualified result claims are what the rules restrict, and they are also what engines discount. Have your compliance reviewer read the new copy before it ships.

Should each attorney have their own page?

Yes. An engine resolves a firm through its people, and a client often asks about a named lawyer rather than a firm. Each attorney page carries the bar number, admission year, courts, a real biography and photo, and a Person node in structured data linking to the bar profile.

Which practice areas should we fix first?

The ones clients ask assistants about with urgency and a place: injury, criminal defense, family, immigration, probate. Start with the two that bring the most calls, rewrite those pages first, and measure them on the baseline questions before touching the rest.

What is the free South Florida AI Search Boost for law firms?

A free, no-obligation read for firms practicing in Palm Beach, Broward and Miami-Dade counties: dated screenshots of what ChatGPT, Gemini, Perplexity and Google AI Overviews say for client questions in your county, a crawler-access check, a read of your practice-area pages, and a written fix list. Email info@suedeai.ai or DM @AISUEDE on X, and the founder replies.

Back toSuede AI SEO