When a prospective client asks ChatGPT for the "best probate attorney in Dallas", the model returns a conversational paragraph naming three to five firms. There is no map pack, no sponsored row, no list of ten blue links to scroll through. The firms it names are the firms that get the inbound call. Everyone else is invisible to that engagement.
The selection is not random and it is not driven by your website's design. It is driven by entity authority — the cross-referenced signal that you are a real, recognised, recommended legal practice. This guide is the playbook for building that authority for a law firm specifically: the schema, the directory triad, the city-by-practice page architecture, and the FAQ content that gets your name into the answer.
Why AI Search Broke the Law Firm Marketing Playbook
The traditional legal SEO contract was simple: rank #1 on Google for "[city] [practice area] attorney", outbid competitors on Google Ads for the same phrase, and convert a percentage of the resulting clicks. Two decades of legal marketing budgets — routinely $50–$300 per click in competitive metros — were built on that single assumption.
ChatGPT and Perplexity do not rank. They cite. When someone types "I need a personal injury lawyer in Houston who handles trucking cases", the model returns a synthesised recommendation naming three to five firms with a one-sentence reason for each. The user does not see ten alternatives. They see the shortlist.
The implication: a Google Ads campaign for "Dallas DUI attorney" no longer captures the search if the prospective client opens ChatGPT first. The auction does not even run. And which firms are named is determined by entity authority signals — Super Lawyers, Best Lawyers, Avvo, Justia, bar association directories — not by historical paid-search performance.
This is not a partial substitution. BrightLocal's 2025 legal-vertical survey found 71% of consumers researching a legal matter consult an AI assistant before contacting any firm, and the share is growing roughly two percentage points per quarter. For practice areas with a higher matter value — estate planning, business litigation, complex injury — the share trends higher still, because the perceived cost of choosing the wrong firm justifies more research.
The economics for the firm side compound: an initial consultation fee in major metros now averages $400 or more, average matter values run $5K–$50K across mainstream practice areas, and an AI citation that produces three to eight qualified consultations per month at the high end of that range is, conservatively, a six-figure annual revenue line. Yet the firm is paying $0 per click for it.
The 5 Sources ChatGPT Trusts Most for Legal Recommendations
BrightLocal's citation analysis of legal-vertical ChatGPT responses identified five sources that appear in the model's reasoning at a meaningfully higher frequency than any others. These are the entities ChatGPT trusts to vouch for a law firm:
"Present" is not enough. A claimed but skeletal profile contributes almost nothing. The threshold is a complete, photo-rich, full-bio profile with current verdicts and settlements where ethically permitted, third-party recognitions, a substantive practice description, and the full attorney roster. ChatGPT distinguishes between "has a listing" and "has a listing that an editor would have approved."
Plus the bar association layer: every state bar maintains a public attorney directory, and most metropolitan areas have a county or city bar association directory in addition. These are not impressive marketing channels, but they are first-tier entity-resolution sources for AI. They confirm that the attorney named in your schema, your Avvo profile, and your Super Lawyers entry is in fact the same licensed person. Without that confirmation, the model treats the entire profile network with reduced confidence.
LegalService Schema — What to Mark Up
For a law firm, use schema.org/LegalService rather than the generic LocalBusiness type. LegalService is a recognised subtype that signals the practice category to AI platforms directly, without requiring them to infer it from your copy. The fields that carry the most weight are areaServed, serviceType, hasOfferCatalog, founder/employee as Person, and knowsAbout.
The example below is for a hypothetical Dallas probate firm, structured the way we deploy it for our legal clients. Replace placeholder values, paste into the <head> of your homepage, and validate at validator.schema.org.
{
"@context": "https://schema.org",
"@type": "LegalService",
"@id": "https://example-probate.com/#firm",
"name": "Caldwell & Reeve Probate Attorneys",
"description": "Probate, estate administration, and trust litigation firm serving Dallas County and the surrounding metroplex since 2008.",
"url": "https://example-probate.com/",
"telephone": "+1-214-555-0142",
"priceRange": "$$$",
"address": {
"@type": "PostalAddress",
"streetAddress": "1700 Pacific Ave, Suite 2200",
"addressLocality": "Dallas",
"addressRegion": "TX",
"postalCode": "75201",
"addressCountry": "US"
},
"areaServed": [
{ "@type": "AdministrativeArea", "name": "Dallas County, TX" },
{ "@type": "AdministrativeArea", "name": "Collin County, TX" },
{ "@type": "City", "name": "Plano" },
{ "@type": "City", "name": "Frisco" }
],
"serviceType": [
"Probate Administration",
"Will Contests",
"Trust Litigation",
"Estate Planning",
"Guardianship"
],
"hasOfferCatalog": {
"@type": "OfferCatalog",
"name": "Probate & Estate Services",
"itemListElement": [
{ "@type": "Offer", "itemOffered": { "@type": "Service", "name": "Uncontested Probate" } },
{ "@type": "Offer", "itemOffered": { "@type": "Service", "name": "Will Contest Representation" } },
{ "@type": "Offer", "itemOffered": { "@type": "Service", "name": "Independent Administration" } }
]
},
"founder": { "@id": "https://example-probate.com/attorneys/m-caldwell#person" },
"knowsAbout": [
"Texas Estates Code",
"Independent vs Dependent Administration",
"Muniment of Title",
"Heirship Determination"
]
}
Why knowsAbout matters disproportionately. It is the most underused property in legal schema and the one that most directly tells AI platforms "this entity has codified expertise in these named topics." For a probate firm, items like "Texas Estates Code" and "Muniment of Title" map onto the exact substrings ChatGPT looks for when a query mentions a specific statute or instrument.
The Lawyer Bio Is the Most Undervalued AI Asset
Most firm websites treat the attorney bio as a decorative HR page — a headshot, a paragraph, a list of practice areas. For AI visibility, the bio is the firm's single highest-leverage content asset. It is where the attorney's credentials, recognitions, and named subject expertise live in the form ChatGPT looks for.
Each attorney needs a Person schema entity, ideally on its own URL, with award, alumniOf, memberOf, knowsAbout, and publishingPrinciples populated. Pair it with a visible, well-written bio on the page so the structured data and the human-readable content reinforce each other.
{
"@context": "https://schema.org",
"@type": "Person",
"@id": "https://example-probate.com/attorneys/m-caldwell#person",
"name": "Margaret Caldwell",
"jobTitle": "Founding Partner",
"worksFor": { "@id": "https://example-probate.com/#firm" },
"alumniOf": [
{ "@type": "CollegeOrUniversity", "name": "University of Texas School of Law", "sameAs": "https://law.utexas.edu/" }
],
"memberOf": [
{ "@type": "Organization", "name": "State Bar of Texas" },
{ "@type": "Organization", "name": "ABA Section of Real Property, Trust & Estate Law" },
{ "@type": "Organization", "name": "Dallas Bar Association, Probate & Trust Section" }
],
"award": [
"Super Lawyers (Texas), Estate & Probate, 2021–2025",
"Best Lawyers in America, Trusts and Estates, 2024–2025",
"Texas Board Certified — Estate Planning & Probate Law (TBLS)"
],
"knowsAbout": [
"Texas Probate",
"Will Contests",
"Fiduciary Litigation",
"Trust Modification"
],
"publishingPrinciples": "https://example-probate.com/editorial-standards"
}
Awards are the highest-signal field. "Super Lawyers (Texas), Estate & Probate, 2021–2025" is a string the model can confidently pattern-match against Super Lawyers' own indexed pages. That cross-reference is what graduates the attorney from "claimed credential" to "verified recognition" in the model's confidence calculus.
Practice Area Pages — the City × Practice Combination
The single most under-built asset on most law firm sites is the city × practice area matrix. The right architecture is one dedicated page per combination. A Dallas-based firm with four practice areas serving six metroplex cities is therefore 24 pages, not four. Each page should be substantive and unique — not boilerplate with the city name swapped.
This is the architecture that wins ChatGPT citations for specific queries: "best estate planning attorney in Plano" rather than the metro-wide "Dallas estate planning attorney" race. The competition at the niche-by-city level is dramatically thinner, and that is where AI sends its earliest and most qualified consultations.
What each city × practice page must contain: a practice area overview written for that city; the relevant state statutes named explicitly (e.g., "Texas Estates Code § 401.003 governs independent administration"); recent representative verdicts or settlements with structured data where ethically permitted; the local court and clerk information (Dallas County Probate Court 1, Collin County Probate Court, etc.); and a city-specific FAQ. This combination is what gets the page treated as the authoritative resource for that exact query.
Build the matrix incrementally. A firm cannot ship 24 high-quality pages in a month, but it can ship two per week for three months and have the full grid live by quarter end. The pages compound: each one shares schema and internal links with the others, and the firm's overall entity authority rises with every page indexed.
FAQ Content — the Highest-Converting AIO Asset for Lawyers
When a prospective client types "how much does an uncontested divorce cost in Texas" into ChatGPT, the FAQ that answers it cleanly — with a real range, a clear explanation of the variables, and a named jurisdiction — is the FAQ ChatGPT cites. The model is looking for self-contained answers to natural-language questions. The firm whose FAQ supplies them gets the engagement.
Pages with properly implemented FAQPage JSON-LD are selected as ChatGPT sources at 2.1× the rate of equivalent pages without it. There is no higher-leverage on-page change available to a law firm.
Eight examples of high-citation legal FAQ prompts — build pages around questions like these:
- How much does an uncontested divorce cost in Texas?
- When do I need a will vs a trust in California?
- What is the average settlement for a rear-end collision in Florida?
- Can I get DUI charges reduced to reckless driving in Georgia?
- How long does probate take in Dallas County?
- What is the statute of limitations for a slip-and-fall in New York?
- Do I need a lawyer for an uncontested guardianship in Arizona?
- How are non-compete agreements enforced in Massachusetts?
Each answer should be 40–80 words, complete in itself, with the firm and jurisdiction named naturally. The combined effect of city × practice area pages plus per-page FAQ blocks is the structural foundation of every well-cited legal site.
The Avvo / Super Lawyers / Best Lawyers Triad
Three recognition platforms carry disproportionate weight in ChatGPT's legal citations. Treat them as the foundation of AI visibility for attorneys; everything else amplifies their effect.
-
Avvo — score 9.0 or higher Claim the profile, add a full bio, add representative cases, add every recognition, and publish 5–10 substantive Legal Guides on your practice areas. Solicit client reviews methodically. An Avvo score of 9.0+ is the threshold at which the profile begins surfacing meaningfully in AI citations.
-
Super Lawyers — engage the nomination process Super Lawyers uses a peer-nomination and independent-research process administered by Thomson Reuters. There is a paid step (the selected listing fee), but selection itself is editorial. Once recognised, the badge appears in your Person schema
awardfield and is the single most-cited credential in ChatGPT's legal responses. -
Best Lawyers — apply through peer review Best Lawyers in America is selected by exhaustive peer-review balloting among practising attorneys. The application is straightforward, but selection requires that peers in your jurisdiction know your work — participating in local bar sections and CLE programming is the practical path in.
None of the three is achievable overnight, and none is purchasable in the way an ad placement is. That is precisely why they carry signal: they are filters AI platforms trust because the entities behind them have institutional reputations to protect.
What You Can and Cannot Say in AI-Optimised Content
State bar advertising rules apply to AI-discoverable content exactly as they apply to any other firm marketing. The rules vary by jurisdiction, but the principles are consistent and easy to honour without weakening the content for AI:
Avoid unqualified superlatives. Do not write "best probate attorney in Dallas" as a self-description. Do write "Recognised by Super Lawyers as a top probate attorney in Texas, 2024." The second is a verifiable third-party attribution and is permissible across every state bar. It also reads as a stronger signal to AI platforms, because the recognition is named and dated.
Two further guardrails: explicitly disclaim the formation of an attorney–client relationship in any FAQ or guide content that could be construed as legal advice; and note "results vary; past results do not guarantee future outcomes" alongside any published verdict or settlement figure. Both disclaimers are short, are required by most bars in some form, and do not reduce the citation value of the surrounding content.
Timeline & ROI
The pipeline from schema deployment to AI-driven consultation requests is predictable. There are no shortcuts that compress the AI-platform learning curve, but the firms that start now begin appearing roughly two months before their later-starting competitors.
Frequently Asked Questions
See where AI stands on your firm right now
Run the free AI Visibility Diagnostic — we check your LegalService schema, attorney Person markup, bar profile completeness, and simulate "best [practice] attorney in [city]" queries across all 7 AI platforms.
The Dominant tier is calibrated for legal-vertical economics — full triad management, monthly city × practice content, and per-attorney bio optimisation.