When a potential client sits down with ChatGPT and asks "who is a good family law attorney in [your city]," they are not doing a Google search. They are having a conversation with a tool that is about to recommend a specific person or firm.
The question is whether that recommendation is you.
This is already happening at meaningful scale. A 2026 BrightLocal survey found that 45% of consumers now use AI tools for local business recommendations — and the legal category, because of its high-trust, research-intensive decision process, is one of the most active query categories. A potential client who asks ChatGPT for an attorney has already decided they need one. They are in the shortlisting stage. Being in that answer is the equivalent of getting a referral — except the referral is coming from a tool that is running on millions of devices simultaneously.
This playbook covers the 6 signals that determine whether AI tools recommend your firm, in the order that matters most.
The 6 signals that determine whether ChatGPT recommends your firm
These are not abstract ranking factors. Each one corresponds to a specific thing AI tools do when they process a query like "best estate planning attorney in [city]." Work through them in order — the earlier ones are foundational.
Signal 1: Your NAP data is consistent everywhere
NAP stands for name, address, and phone number — but the concept extends to your website URL and the exact formatting of your firm name. AI tools cross-reference your information across multiple sources before recommending you. Inconsistencies between those sources reduce their confidence.
For a law firm, NAP inconsistencies are common in specific ways:
- The firm name uses "Law Office of [Name]" on the website but "Law Offices of [Name]" on Avvo
- The address changed when you moved offices and was updated on your website but not on Justia or FindLaw
- The phone number that appears on Yelp forwards to an old line that was transferred to a new number
None of these feel serious when you are managing a practice. But each one is a discrepancy that reduces an AI tool's willingness to describe you with confidence.
What to audit: Compare your firm name, address, phone number, and website URL across: your website footer, Google Business Profile, Yelp, Avvo, Justia, FindLaw, LinkedIn, and your state bar's directory listing. Start with those eight. Fix any mismatches to match your website exactly — same punctuation, same spacing, same abbreviations.
Signal 2: Avvo, Justia, and FindLaw profiles are complete
BrightLocal's research into how AI tools source legal recommendations consistently finds these three directories cited in legal query responses. They are the legal-specific corroboration layer that Google Business Profile and Yelp cannot provide — they signal to AI tools that your firm exists, is legitimately practicing, and is what it claims to be.
An incomplete profile is almost as problematic as a missing one. An Avvo profile with no practice areas listed, no biography, and no peer endorsements gives AI tools almost nothing to work with. A complete Avvo profile — with every practice area checked, a clear attorney biography, a client rating, and peer endorsements — is a rich information source that AI tools can cite and summarize.
What to do: Log into each directory and treat it as a mini-website for your firm. For each one:
- Confirm your firm name, address, and phone number match your website exactly
- Complete all practice area selections — be specific, not just "litigation" or "family law"
- Write a biography that describes who you help and what you help them with (not just credentials)
- For Avvo specifically: complete the peer endorsement requests if you have not already — attorney endorsements carry weight in Avvo's internal authority model, which flows through to AI citations
Signal 3: Each practice area has its own specific page
This is where most law firm websites fail. A practice areas page that lists "Family Law, Estate Planning, Business Formation, Personal Injury" with two sentences under each is useful for navigation, but it gives AI tools almost nothing to match against a specific client query.
When someone asks ChatGPT for "an estate planning attorney who handles living trusts and powers of attorney in [city]," the tool is looking for a page that specifically addresses living trusts, powers of attorney, and the relevant city. A generic practice areas page with "We handle estate planning" does not satisfy that query with enough confidence to generate a recommendation.
The law firms that consistently appear in AI recommendations have individual pages for each meaningful practice area — with content that answers the questions a real potential client would have: What does this involve? Who is it for? What does it cost, approximately? What happens if I do not have this in place?
What to build: For each practice area you want to be recommended for, create a dedicated page that includes:
- The specific legal matter named in the H1 (e.g., "Estate Planning Attorney in [City]" not "Estate Planning Services")
- 2–3 sentences on who typically needs this service — lead with this, not with your credentials
- What the engagement looks like in plain language — what happens after someone calls you
- At least 3–5 questions that clients commonly ask, written out as visible on-page Q&A (not hidden in an accordion), each answered in 2–4 direct sentences
- Your city or service area mentioned naturally at least twice in the body text
The Q&A sections deserve specific attention. AI tools frequently pull question-and-answer content verbatim when answering a query that matches one of your questions. A page that has "What does an estate plan include?" followed by a specific, accurate four-sentence answer is giving ChatGPT something it can cite directly when a potential client asks that exact question. Write those questions the way a real client would ask them — not "What are the components of a comprehensive estate planning document?" but "What do I actually get when I hire an estate planning attorney?"
This is the single highest-leverage change most law firms can make. One well-structured practice-area page can shift an AI recommendation.
Signal 4: Your schema markup covers the firm and individual attorneys
Schema markup is the machine-readable layer of your website — it tells AI tools directly what your business is, what it does, and who practices there, in a format that requires no inference.
For a law firm the minimum schema you need is:
LegalServiceschema (a subtype ofLocalBusiness) on your homepage with your firm name, address, phone number, service area, and the legal practice areas you coverFAQPageschema on each practice-area page — the questions and answers you write for Signal 3 can be marked up directlyPersonschema for each attorney, linked from their bio page, with their name, professional credentials, practice areas, and a link to the firm
The Person schema for individual attorneys is particularly important for AI tools that receive queries like "who is [attorney name]" or "what kind of cases does [attorney name] handle." Without it, the AI has to synthesize that information from unstructured text, which it does less reliably.
How to check: Paste your homepage URL into schema.org's validator. Look at what it returns. If you see a very thin LocalBusiness block or nothing at all, that is your gap. A WordPress plugin or a developer can implement basic LegalService and Person schema in an afternoon.
Signal 5: You have a credible review presence on multiple platforms
Reviews are not just a Google thing for law firms. AI tools that answer "who is the best estate planning attorney in [city]" are drawing on review signals from multiple platforms — and for legal, that means Avvo client reviews, Yelp reviews, and Google reviews all contribute to the picture.
A firm with 4.9 stars and 60 reviews on Google but nothing on Avvo and a sparse Yelp listing is less robustly corroborated than a competitor with 30 Google reviews, 8 Avvo client reviews, and a complete Yelp profile with a handful of ratings. The AI is looking for consensus across sources.
Avvo client reviews are particularly underused by attorneys. They are harder to collect than Google reviews because clients have to create an Avvo account, but they carry specific weight in legal AI recommendations because of Avvo's position as a legal-specific authority.
What to do: After each matter closes — or at any natural touchpoint with a satisfied client — send a direct link to your Avvo review page alongside your Google review request. Do not ask directly in client communications (check your state bar's rules on this), but a follow-up email that says "If you're able to share feedback, it helps others facing similar situations find the right help" typically works well.
Signal 6: AI crawlers can reach your website and your practice-area pages
This is the baseline check that should have been done first but often gets forgotten. If GPTBot, ClaudeBot, or PerplexityBot are blocked in your robots.txt file — often as collateral damage from a security plugin or a past technical change — AI tools cannot read your site at all. Your practice-area pages, schema, and service descriptions are invisible to them regardless of how well written they are.
Check: Open yourdomain.com/robots.txt in a browser. Look for any Disallow: / lines, or any specific disallows for GPTBot, ClaudeBot, or PerplexityBot. If you find them, that is an urgent fix. This five-minute check can unblock months of invisible AI credit for work you have already done.
What a well-optimized law firm looks like to ChatGPT
To make this concrete — here is the kind of description ChatGPT produces for a law firm that has done this work, versus one that has not.
Firm A (signals in place): "[Firm name] is a family law practice in [city] specializing in high-conflict custody cases and divorce proceedings. The firm is led by [attorney name], who has 12 years of experience in [state] family law. They offer free 30-minute consultations. Their Avvo rating is 9.2 with 14 client reviews. You can reach them at [phone number]."
Firm B (signals missing): "I found what appears to be a family law practice in your area, but I don't have detailed information about their specific services or reputation. You may want to verify directly."
Both firms exist. One has given AI tools a complete, consistent, specific picture. The other has not. The recommendation goes to Firm A every time.
The legal-specific directory checklist
These are the profiles worth claiming and completing in priority order. Each one is a citation source that AI tools actively pull from when responding to legal queries.
| Directory | Priority | Why it matters for AI |
|---|---|---|
| Google Business Profile | Critical | Anchors local trust for all AI tools |
| Avvo | Critical | Legal-specific authority, cited in ChatGPT and Perplexity responses |
| Justia | Critical | Legal-specific index, frequently cited by AI for attorney profiles |
| FindLaw | High | Attorney directory with high domain authority and AI citation frequency |
| Super Lawyers | High | Selection-based recognition that reads as a third-party endorsement |
| State bar directory | High | Official corroboration — the most authoritative legal citation available |
| Yelp | High | Review platform AI uses for consumer trust signals |
| LinkedIn (firm and individual) | Medium | Practitioner-level verification, Person entity corroboration |
| Martindale-Hubbell | Medium | Long-established legal directory, Avvo and Martindale are the same company |
| AVVO Lawyer Directory | Critical | Already listed — do not duplicate |
Where to start this week
If you are starting from zero on this, here is the order that produces results fastest:
- Check robots.txt — five minutes, potentially the most impactful fix
- Audit NAP across eight core platforms — standardize everything to match your website
- Claim and complete Avvo, Justia, and FindLaw profiles — if not already done
- Validate your homepage schema — use schema.org's validator
- Pick your highest-value practice area and rewrite that page — add specificity, city, and a FAQ section
- Request 2–3 Avvo client reviews from past clients — slow build, start now
Items 1–3 are a matter of hours. Items 4–6 take more time but also produce more sustained AI visibility. The full list is a one-time project — once the foundation is in place, maintaining it is mostly keeping information current.
The realistic picture
AI search for legal queries is not yet replacing referrals or replacing Google as a lead source. The realistic expectation is that over the next 6–18 months, a proportion of clients who used to start their research on Google will start it on ChatGPT — and those clients will typically arrive at your intake form warmer and more pre-qualified than a cold organic click, because an AI tool has already told them you are the right fit.
That shift is already underway. A San Francisco attorney became a $19,000 client for an agency that appeared in ChatGPT — self-reported through intake. A home inspection business owner reports that every month, more callers mention finding the company on ChatGPT. The pattern is consistent across professional services.
The law firms that close these gaps now will be established in AI answers when inquiry volume from this channel becomes significant. The ones that wait will face the same problem you are facing today with your competitor — wondering why the AI keeps recommending them instead.
If you want to know where your firm currently stands before you start, the LovedByAI GEO audit scans for the signals covered in this playbook — bot access, schema completeness, NAP consistency, service page depth — and returns a prioritized fix list specific to your site. It takes about two minutes.
