A prospect opens ChatGPT, types “best injury lawyer near me,” and reads the answer. If an AI answer leaves your firm out, a prospective client may never consider you as an option. This isn’t about rankings or traffic. It’s about whether AI hands your name to someone ready to hire. In a 2026 survey of 1,110 US consumers, half said they would use at least one AI engine to research a lawyer. This guide gives you five steps to get cited, measure it, and connect it to signed cases. You can start the first one this week.
Why clients now ask AI before they search Google
Clients ask AI first because it answers the question instead of handing them ten blue links to sort through. In iLawyerMarketing’s 2026 survey of 1,110 US consumers, 41.9% said they would use ChatGPT to research which lawyer to hire, up from 9% in 2023, and 50.1% said they would use at least one AI engine. Google still leads for attorney research, but its share fell to 71.9% in the same survey, down from 86.7% a year earlier. The pattern looks like “ask AI first, then verify on Google,” since over 70% of ChatGPT users said they would still use Google too. That shift matters because fewer people click through at all. Pew Research Center, analyzing 68,879 Google searches in March 2025, found that when an AI summary appeared, users clicked a traditional result just 8% of the time.
How AI engines decide which law firms to cite
AI answer engines don’t rely on traditional search rankings alone. They retrieve information and generate answers, with sources and citation behavior varying by platform. So the question isn’t “Do I rank?” It’s “Am I the kind of source an AI will retrieve and name?” Engines behave very differently, too. In a 2026 University of Toronto study of 158,099 answers, ChatGPT cited about 2 pages per answer, Gemini about 10, while Perplexity returned no citations in roughly 96% of the answers examined. (Chen and Koudas).
The Toronto study found that FAQ blocks, schema markup, and author or summary boxes were not associated with more citations on the same site. Retrieval by the engine’s search system mattered more. Page titles and paragraphs that match the query also appeared to help. Separately, a controlled Princeton experiment on general web content found that adding statistics, quotations, and cited sources raised a page’s visibility in AI answers by up to about 40% (Aggarwal et al., 2024). That’s a general-web result, not a law-firm study, so treat it as direction. A 2026 qualitative audit by 5WPR and the Haute Lawyer Network found that legal directories, including Chambers, Legal 500, Super Lawyers, Best Lawyers, Martindale-Hubbell, Avvo, and Justia, were among the sources examined. The audit covered three query types across eight practice areas and tested firm-specific queries against three Am Law 100 firms, so its findings should not be treated as a market-wide ranking.
Step 1: Audit your firm’s current AI visibility
Start by finding out what the engines say about you today. Run a fixed set of prompts across ChatGPT, Perplexity, Claude, and Google AI Mode, logged out, with your location set to your market. Here’s a 10-prompt audit you can run this week.
The 10-prompt audit you can run this week
| Prompt category | Example prompt | What to record |
| Category and city superlative | “Best personal injury lawyer in [city]?” | Is your firm named? Cited with a link? Which firms and sources appear? |
| Cost | “How much does a [practice] lawyer cost in [city]?” | Whether you’re named, and which sources the answer leans on |
| Situational advice | “I was just in a car accident in [city], what do I do?” | Whether any firm is named, and whether yours surfaces |
| Direct reputation check | “Is [your firm] any good?” | What the engine says about you and where it pulls from |
| Entity comparison | “[Your firm] vs [competitor] for [matter]?” | How you’re framed against named competitors |
| Proximity | “[practice] lawyer near [neighborhood]” | Whether you appear for local intent |
| Process guidance | “How does a [matter] case work in [state]?” | Whether your pages are cited as the explainer |
| Specialty niche | “Lawyer for [narrow matter] in [city]” | Whether you surface on narrow intent |
| Outcome question | “Typical settlement for [matter] in [state]?” | Whether you’re cited, and whether it’s an answer you could repeat within bar rules |
| Aggregated recommendation | “Give me three [practice] firms in [city]” | Whether you’re on the list, and in what position |
Score every answer: Named, Cited, or Invisible
| Metric | What it tells you | Where to track it |
| Named rate | Share of prompts where the answer says your firm’s name | Your prompt spreadsheet |
| Cited rate | Share where your URL is linked as a source | Same sheet, citation-URL column |
| Invisible rate | Share where you’re neither named nor cited | Same sheet |
| Qualified leads | Inbound that matches your case criteria | Intake log or CRM |
| Signed cases | Clients you actually retained | Case management system |
Step 2: Make your firm a machine-readable entity
Before an engine can name you, it has to recognize you as one consistent entity. Keep your name, address, and phone number identical across your site, your Google Business Profile, and every directory. Claim and fill out the Google Business Profile completely. Add schema markup (Organization, LegalService, Attorney, LocalBusiness) so your details are machine-readable. The Toronto study found schema and similar markup don’t, on their own, win more citations. Do it for recognition and consistency. It’s the groundwork that lets an engine connect your site, your profile, and your directory listings into one firm, not a shortcut to the answer box.
Step 3: Build pages AI can quote
AI may cite passages, instead of whole pages, so write pages it can lift. Put the real question in the heading and answer it in the first sentence, because the one on-page factor that held up in the research was how well a page’s title and opening actually match the question asked. Back claims with statistics and name your sources, which is what the Princeton experiment rewarded. Add FAQ schema to the questions clients ask. It helps machines parse the page; it doesn’t buy you a citation. Rebuild one deep practice-area page and its matching city page at a time, rather than spreading a thin rewrite across fifty pages that moves nothing.
Step 4: Earn the authority signals that get you named
Engines lean on outside proof more than your own copy, so build it in three places. Reviews: ask every satisfied client, and keep them current. Legal directories: the ones AI already pulls from, named above, are worth claiming and completing. Digital PR: coverage and mentions on third-party sites the engines treat as credible. For a firm the model doesn’t already know, the Toronto study found a cited third-party page that names you is one of the most important things that gets you named at all.
Step 5: Measure, iterate, and tie AI visibility to signed cases
Named and cited rates are the start. Follow the chain all the way down: named, then cited, then qualified leads, then signed cases. If your cited rate climbs but signed cases don’t move, you’re winning the wrong prompts. Re-run the audit monthly, keep the changes that produce cases, and drop the ones that only produce mentions.
Where AI citation work fits with your existing SEO
AI visibility isn’t a replacement for SEO, and it isn’t the same job. The Toronto study found that the pages engines retrieve and cite are mostly not the ones ranking at the top of Google or Bing for the same query, so a page-one ranking is not a stand-in for getting cited. That said, the two still feed each other. Keep your site crawlable, fast, and properly marked up, then build the useful content and third-party mentions that can help AI engines identify your firm.
Stay compliant while you optimize
Everything you publish to get cited is still lawyer advertising, so the ethics rules apply. ABA Model Rule 7.1 bars false or misleading claims about your services, and that covers your website, not just your billboards. If you publish fees to answer cost prompts, many states require you to disclose any additional costs, and a contingency-fee claim has to say who pays expenses. Reviews count too: the FTC’s Endorsement Guides may also apply to reviews and testimonials used in marketing. State bar rules vary and override any general rule of thumb, so check yours before you publish fees, case results, or testimonials.
Frequently asked questions
How long does it take for Google AI Mode to start citing law firm?
There’s no fixed timeline. AI answers shift from week to week, and the Toronto study found the set of pages an engine cites drifts steadily over a matter of weeks. A citation also depends on the groundwork landing first: entity consistency, pages worth quoting, and outside authority signals. Treat it like SEO: months of steady work rather than a switch you flip, and re-run your audit monthly so you can see the movement.
Can my firm guarantee an AI answer engine will recommend it?
No, and be skeptical of anyone who promises it. The engines are selective. SOCi’s 2026 Local Visibility Index, looking at roughly 350,000 locations (not specifically law firms), found ChatGPT recommended just 1.2% of locations, versus 35.9% appearing in Google’s local pack. You can improve your odds with the steps above. You cannot buy a guarantee.
Do I still need legal directories if AI mostly cites firm websites?
Yes. The 2026 5WPR and Haute Lawyer Network audit found that directories such as Super Lawyers, Avvo, Martindale-Hubbell, and Justia featured prominently in the AI citation results it examined. Keep your profiles accurate and complete, but remember that the audit was limited in scope.
Which practice area pages should I rebuild first?
Start where a signed case is worth the most and real demand exists. Rebuild one high-value practice-area page and its city page, confirm the method produces cases, then repeat. Rebuilding everything at once spreads your effort too thin to learn anything.
How do I track my firm gets mentioned in AI answers without buying software?
Build a prompt list, run it logged out across ChatGPT, Perplexity, Claude, and Google AI Mode, and log three things per answer in a spreadsheet: were you named, were you cited with a link, and in what position. Re-run it monthly. That’s the free version of what the paid tools automate.