Lawyer AI Overviews Citations — the YMYL Playbook

The work is concentrated where the YMYL gravity is strongest. Attorney bios do most of the heavy lifting because Google’s rater guidelines¹ treat YMYL pages with anonymous or thin authorship as low quality almost regardless of content. Statute-citation discipline does the second-largest share because it converts the firm’s legal pages from “blog posts about law” into “reference material with sources” — which is what AI engines preferentially lift from.

Google’s quality rater guidelines¹ describe Trust as “the most important member of the E-E-A-T family” and add explicitly that “an untrustworthy page has low Page Quality regardless of how Experienced, Expert, or Authoritative it may seem.” Search Engine Land’s 2026 GEO research² finds the same pattern in AI citation behaviour: engines preferentially cite sources with named authors, verifiable expertise, and inline source attribution — and the preference compounds on YMYL queries.

The implication for a boutique firm is asymmetric. A photographer’s site with a thin About page still earns citations on photography queries; a law firm site with a thin Attorneys page does not earn citations on legal queries even if the rest of the site is meticulously written. The YMYL gravity makes the attorney-bio investment the highest-leverage work the firm does for AI search visibility — higher than schema, higher than content production, higher than directory presence.

The realistic 2026 ambition for a boutique firm is to be cited on the procedural and informational legal queries within the firm’s practice area and jurisdiction. “Statute of limitations medical malpractice california” produces an AI Overview that cites two to four sources; if the firm’s page on the topic is well-written, cites California Code of Civil Procedure § 340.5 by name, and carries a Person JSON-LD pointing to a real California-admitted attorney with relevant practice experience, the page is competitive for one of those citation slots. The traffic that follows is small but high-intent: someone researching the procedural question becomes a candidate client by the time they reach the firm’s contact form.

The hiring queries are largely unreachable in AI search and the firm should not optimise for them. ChatGPT and AI Overviews explicitly decline to recommend a specific attorney on most “best lawyer for X” queries — they route the user to the local pack, to Avvo, to Justia, or to a state bar referral service. Optimising the Squarespace site for “best estate planning attorney” through the lens of AI citation is a misallocation of effort; that ranking lives in the local pack and the directory layer, which the local SEO leaf handles.

The pattern is testable. Search ChatGPT for “statute of limitations medical malpractice california” and inspect the citations. The top citation is typically Cornell LII or Justia, both of which carry the statute text directly. The second and third citations split between state bar publications, law school clinics, and boutique-firm pages that cite the statute by number and add procedural detail (the discovery rule, the cap on non-economic damages under California Civil Code § 3333.2 (Medical Injury Compensation Reform Act, MICRA), the one-year-from-discovery deadline). The pages that lose are the ones whose paragraphs assert the rule without citing it.

For a boutique Squarespace site, the implication is direct. Every legal information page on the firm’s site should cite its primary sources inline. If the page describes the California small estate threshold, it cites California Probate Code § 13100. If the page describes the New York statute of limitations on contract claims, it cites N.Y. C.P.L.R. § 213. If the page describes a Texas procedural rule, it cites the Texas Rule of Civil Procedure. The citation discipline is identical to a brief — the firm’s public-facing legal pages should look like the firm’s actual work.

The pattern below is the production block this site recommends for a single-attorney Person entity. The full multi-attorney pattern, including how the attorney Person joins to the LegalService firm entity through the employee property, lives in the lawyer schema leaf. Schema.org’s Person specification³ describes knowsAbout as “indicates topics a person has knowledge about, suggesting possible expertise but not implying it” — the wording matters because knowsAbout is a self-asserted signal that AI engines weight more heavily when the same expertise is also verifiable through external sources.

 <script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Person", "name": "Anjali Hartwell", "url": "https://hartwellcole.com/attorneys/anjali-hartwell/", "jobTitle": "Founding Partner, Hartwell & Cole, LLP", "image": "https://hartwellcole.com/assets/anjali-hartwell.jpg", "knowsAbout": [ "Estate Planning", "Trust Administration", "Probate Litigation", "California Probate Code" ], "alumniOf": { "@type": "EducationalOrganization", "name": "UC Berkeley School of Law" }, "memberOf": [ { "@type": "Organization", "name": "State Bar of California" }, { "@type": "Organization", "name": "American Bar Association, Real Property, Trust and Estate Law Section" } ], "sameAs": [ "https://www.avvo.com/attorneys/example", "https://www.justia.com/lawyers/example", "https://www.martindale.com/attorney/example", "https://www.linkedin.com/in/example", "https://apps.calbar.ca.gov/attorney/Licensee/Detail/000000" ] } </script> 

The state bar attorney-lookup URL in the sameAs array does specific work here. AI engines treat bar association attorney lookups as primary verification: a Person whose sameAs includes the State Bar of California’s public lookup for the named attorney is dramatically harder to confuse with a different attorney of the same name, and the bar lookup itself is one of the strongest YMYL trust signals available.

The implementation on Squarespace is straightforward. Block-level citations live in regular text blocks; inline citations live in the surrounding paragraph. The rendering does not need to match a brief’s typography — AI engines parse plain-text statute references reliably (“California Probate Code § 13100” or “Cal. Prob. Code § 13100” both work) and human readers prefer the un-italicised plain-text format on web pages. The discipline is in the citation, not in the typography.

The work compounds with the practice area pages leaf. Practice-area pages that cite the governing statute, the leading case, and the applicable procedural rule for the firm’s jurisdiction are the pages AI engines lift from preferentially over generic “estate planning attorney” pages from larger competitors whose content is shallower.

The practical 2026 rule for boutique firms: do not publish client testimonials on the Squarespace site in jurisdictions whose bar rules treat testimonials as restricted communications, unless the firm’s ethics counsel has cleared the specific testimonial and the disclaimer language. ABA Model Rule 7.2⁶ permits testimonials in advertising under defined conditions; the state variants are stricter and they vary materially. Florida’s rules in particular have a long history of pre-filing requirements that catch out-of-state firms by surprise. Where the firm cannot or will not vet a testimonial through ethics counsel, the safer pattern is to omit it from the site entirely and let third-party reviews on the directory profiles (Avvo, Google, Justia) carry the social-proof load.

The corresponding AI-search implication is that case-result claims and outcome statistics have to clear the same bar. “Recovered $4.2 million for a single client last year” can be a misleading communication under Model Rule 7.1⁵ if the recovery is presented in a way that suggests typical client outcomes. The conservative pattern is to omit aggregate results claims entirely or to include them with a clear disclaimer that past results do not guarantee future outcomes. The disclaimer language varies by jurisdiction and the firm’s ethics counsel should review the exact wording.