"Best personal injury lawyer in [city]."
Six months ago that query went to Google. The person scrolled through ads, clicked a few blue links, bounced between websites, and eventually called someone. Today, 73% of people looking for professional services check an AI engine first. They ask ChatGPT, Perplexity, or Google's AI Overview. And the AI gives them one name. Maybe two.
If your firm is not one of those names, you are not in the consideration set. You are not second or third. You are invisible. You can check if your firm shows up in ChatGPT right now and see for yourself.
Why legal is high-stakes for AI search
Legal is a YMYL category -- Your Money Your Life. Google invented that classification, and every major AI engine has adopted it. It means AI models apply the strictest trust standards when answering legal questions. They will not casually recommend a law firm the way they might suggest a coffee shop.
The trust framework they use is E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness. For legal queries, every one of those signals gets weighted heavily. A firm that demonstrates E-E-A-T through structured data and authoritative content gets cited. A firm that just has a nice website with stock photos and vague practice area pages does not.
This is the core problem. Most law firms built their websites for humans browsing. They did not build them for AI models parsing. The model cannot tell that your managing partner has 25 years of trial experience unless that fact is encoded in a way the model can read -- which means structured data, not a paragraph buried on your about page.
What AI engines look for in a law firm
When someone asks an AI for a lawyer recommendation, the model pulls from a specific set of signals. These are not mysterious. They are documented, measurable, and deployable. Here is what AEO (Answer Engine Optimization) requires for legal practices.
The foundation. LegalService schema tells AI models that your business is a law firm, what practice areas you cover, which jurisdictions you operate in, and where you are located. Attorney schema (using the Person type with relevant properties) registers each lawyer as a distinct entity with bar admissions, years of experience, education, and areas of specialization. Without this, the model has no structured way to understand what your firm does or who works there.
People do not ask AI generic questions. They ask specific ones. "How much is my personal injury case worth?" "What happens after I file for divorce?" "How long does a custody case take in Texas?" FAQPage schema feeds the model answer-ready content in the exact format it synthesizes responses from. If your firm has structured answers to these questions, the AI will cite you when someone asks them. If you do not, it will cite whoever does.
AI models build trust through entity verification. When your Organization schema links to your Avvo profile, your Martindale-Hubbell listing, your state bar directory page, and your LinkedIn company page via sameAs properties, the model can cross-reference your firm across multiple authoritative sources. That cross-referencing is how it decides you are real, credible, and worth recommending. Firms without these connections are entities the model cannot verify -- and unverified entities do not get cited.
Case studies. Legal guides. Published articles in bar journals or legal publications. Content that demonstrates genuine experience with real legal matters. The model is looking for depth and specificity, not keyword-stuffed blog posts about "5 things to know after a car accident." One detailed guide about how contingency fee structures work in your state is worth more than fifty thin articles written for search crawlers.
Google reviews, Avvo ratings, Lawyers.com reviews -- these feed directly into the trust calculation. The model aggregates review data from multiple platforms. A firm with 47 Google reviews averaging 4.8 stars and a 10.0 Avvo rating is a firm the model can confidently recommend. A firm with three reviews and no Avvo profile is a risk the model will not take, especially in a YMYL category.
Practice area specifics matter
Not every law firm needs the same schema. The signals that drive AI citations for a personal injury firm are different from the ones that work for estate planning. This is where most generic marketing agencies fail -- they apply the same template to every practice area and wonder why nothing happens.
Personal Injury
Case result data, settlement ranges, contingency fee structure, injury-specific FAQ schema
Family Law
Process guides (divorce timeline, custody factors), jurisdiction-specific FAQ, mediator credentials
Estate Planning
FAQ about wills, trusts, probate process, elder law credentials, state-specific estate tax content
Criminal Defense
Rights and process content, charge-specific FAQ, court procedure guides, former prosecutor experience
A PI firm that publishes structured case result data -- average settlement by injury type, timeline from filing to resolution -- gives the AI model exactly the kind of specific, authoritative content it needs to cite that firm when someone asks "how much is a broken leg case worth." A family law firm that publishes a detailed guide to custody factors in their state, marked up with proper Article schema and FAQ schema, becomes the go-to citation for custody questions in that jurisdiction.
The trust multiplier
Here is what makes legal AEO different from legal SEO. In traditional search, a blue-link click is low-commitment. The person clicks, scans the page, and usually bounces. Conversion rates for law firm websites from organic search hover around 2-4%.
An AI citation is fundamentally different. When ChatGPT tells someone "Contact [Firm Name] for your personal injury case -- they handle cases in [your state] and have a strong track record with [injury type] claims," that is not a link on a page. That is a direct recommendation from a platform the user already trusts. The AI is explicitly telling the person: this is who you should call.
That recommendation carries the weight of the platform's credibility. It is the most qualified lead a law firm can get. The person already knows your name, your practice area, and why you are relevant to their situation before they ever visit your website.
The math on this is simple
The average personal injury case value to the firm is roughly $15,000 (that is a 33% contingency fee on a $40,000-$55,000 settlement). One additional case per month from AI search is $180,000 per year in new revenue.
Family law averages around $11,000 per case. Estate planning averages $3,000-$5,000. Criminal defense ranges widely but often lands in the $5,000-$15,000 range per case depending on charge severity.
Even one extra case per quarter from AI citations pays for an AEO investment many times over. The $500 audit is a rounding error against the value of a single new client.
And unlike paid ads, AEO compounds. The structured data you deploy today continues to feed AI models for months and years. Every month builds on the last. Six months from now, your firm is not just cited once -- it is the default answer for your practice area in your market.
Your competitors are already doing this
The firms that move first on AEO in a given market lock in a compounding advantage. If you are wondering why ChatGPT recommends your competitor instead of you, this is the reason. AI models develop citation patterns. Once a model learns that your firm is the authoritative entity for personal injury in Dallas, displacing you requires the competitor to build more structured data, more authoritative content, and more entity signals than you already have.
Right now, most markets are still wide open. Most law firms have zero structured data. No FAQ schema. No entity linking. No LegalService markup. The bar for getting cited is low because almost nobody is doing this work. That will not last. The firms that deploy AEO in 2026 will own their markets. The ones that wait until 2027 will pay more to catch up and may never close the gap.
Find out where your firm stands
We built a free Schema and AEO Health Check that grades any URL on the signals AI models actually use. No signup. No email gate. Takes 30 seconds. Run your firm's homepage through it and see exactly what is missing.
If you want the full picture, the $500 audit maps every gap across your entire site -- every practice area page, every attorney bio, every piece of content -- and delivers a 90-day roadmap for compound AI visibility.
The question is not whether AI search will reshape how people find lawyers. It already has. The question is whether your firm shows up when they ask.