Legal Services AEO Case Study: How Law Firms Win AI Overviews and ChatGPT Citations

⚠️ Composite case study — synthesized from public patterns; not a verified single-company case.

TL;DR

A 22-attorney US employment-law firm rebuilt 38 practice-area and FAQ pages around answer-first writing, layered FAQ, Attorney, and LegalService schema, and produced 12 representative-matter case studies. In 5 months, AI Overview citations on tracked employment-law queries grew from 7 to 22 (3.1x) and ChatGPT cited the firm in 14 of 60 monitored queries (up from zero). Average answer-paragraph length on practice pages dropped from 312 to 78 words, mirroring the snippet length AI engines extract. The playbook below is reproducible by any mid-size firm in a regulated vertical.

Why legal is uniquely affected by AI search

Legal queries are informational, jurisdiction-sensitive, and detail-rich — exactly the shape Google AI Overviews and ChatGPT prefer to synthesize over a list of blue links. Industry coverage shows AI Overviews trigger more often on legal queries than on average commercial queries, and 41% of consumers now begin their search for a lawyer with an AI assistant rather than a search engine.

This case study tracks what one firm did, what it measured, and what every line item cost in editorial effort — so other firms can pattern-match without re-running the experiment from scratch. For background on why answer-engine optimization differs from traditional SEO, see our Answer Engine Optimization Guide and the Google AI Overviews Optimization Guide.

Firm profile (anonymized)

• Vertical: Plaintiff-side employment law (wage-and-hour, discrimination, wrongful termination)
• Size: 22 attorneys, 4 office locations across 2 US states
• Pre-engagement traffic: ~48,000 organic sessions/month
• Pre-engagement AI visibility: 7 AI Overview citations across 60 tracked queries; 0 ChatGPT mentions; 1 Perplexity citation
• Compliance constraint: All rewrites had to clear an internal ethics review against ABA Model Rule 7.1 (no false or misleading communication) and state-bar advertising rules

The firm already ranked top 10 on Google for most of its tracked queries, which matters because more than 90% of AI Overview citations come from domains that already rank in the top 10 organic results. AEO is therefore additive to existing SEO foundations, not a replacement for them.

Baseline measurement

Before any content changes, the firm captured a citation baseline across three engines using a query panel that mirrors how prospective clients ask employment-law questions:

• 60 monitored queries spread across 12 categories (overtime pay, severance, retaliation, FMLA, EEOC charges, wrongful-termination thresholds, etc.)
• Each query was run weekly on Google AI Overviews, ChatGPT (web-enabled), and Perplexity
• Citations were logged manually plus cross-checked with an AI visibility platform

This matches industry observations: most law firms with strong traditional SEO still under-perform in AI answers because their pages are written for human browsing, not extraction.

The 5-month rebuild plan

Phase 1 — Audit and prioritization (weeks 1-2)

The team picked 38 pages for rewrite, prioritized by:

1. Pages already ranking 4-10 organically for AEO-favorable queries (highest leverage; small jumps in extraction-friendliness produce visible AI citations)
2. Practice-area hubs that would anchor cluster authority
3. FAQ pages because FAQPage schema is consistently cited in AI Overviews at higher rates than unstructured pages

Deprioritized: bio fluff, awards pages, generic blog posts. AEO rewards extractable answers, not marketing copy.

Phase 2 — Answer-first rewrite (weeks 3-10)

Every rewritten page now follows the same structural pattern:

1. H1 phrased as the user's question or the entity (e.g., "Wrongful Termination in California: What Counts and What You Can Recover")
2. A 2-4 sentence direct-answer paragraph immediately under the H1 (target 60-90 words, the extractable range observed in AI Overviews on legal queries)
3. Subhead clusters mirroring the long-tail phrasings of the same question ("Is at-will employment an absolute defense?", "What is constructive discharge?")
4. Each subhead followed by an answer-shaped paragraph, then expansion
5. A representative-matter section linking to anonymized case studies
6. A FAQ block of 5-8 questions with FAQPage schema

Length dropped, but extractability rose. Average lead paragraph length went from 312 words to 78 words. Total page word count stayed roughly flat (1,400-1,800 words) because expansion sections and FAQs absorbed the depth.

Phase 3 — Schema and entity layer (weeks 6-12, parallel)

The firm deployed four schema types per page where applicable:

• FAQPage — every FAQ block, every recurring question subhead with a direct answer
• LegalService — practice-area pages, with serviceType, areaServed, and provider linked to the firm's Organization entity
• Attorney (Person) — every attorney bio, including bar admissions (hasCredential), education, languages, and knowsAbout for practice areas
• Article — long-form guides, with author, dateModified, and reviewedBy

For implementation patterns, see our FAQ Schema for AEO: Implementation Guide and Schema.org for AI Search: Property Reference. Pages with FAQPage schema are cited in AI Overviews at materially higher rates than pages without it.

Phase 4 — Representative-matter case studies (weeks 8-18)

The firm published 12 anonymized case studies ("Representative Matters") of 350-600 words each. Each follows a strict template:

• Client situation (no PII; jurisdiction generalized)
• Legal question presented
• Strategy and statute cited
• Outcome (settlement range, not exact number, for ethics compliance)
• Plain-language takeaway answering "What does this mean for someone in a similar situation?"

This matters for AEO because LLMs reward grounded, specific examples over generic marketing claims. Attorney at Law Magazine notes that AI engines skip past pages that bury the answer under "500 words of firm marketing."

Phase 5 — Authority and entity signals (ongoing)

In parallel, the firm:

• Cleaned up directory listings (Avvo, Justia, FindLaw, state-bar profiles) so name/address/phone and practice-area strings matched the LegalService schema exactly
• Earned 9 trade-press placements (state-bar journal, regional legal news) that linked back to specific practice-area pages, not the homepage
• Added an Organization-level sameAs array linking the firm's domain to its LinkedIn, Bar Association profile, and Crunchbase entity

This mirrors the entity-cleanup + authority-corroboration pattern AI Search Engineers documented across multiple law firm engagements: structured data deployment plus targeted authority signal engineering produced AI visibility where pure SEO had not.

Results after 5 months

Additional behavioral data:

• Average lead-paragraph length on rewritten pages: 312 → 78 words
• Pages with FAQPage schema: 4 → 38
• Anonymized representative-matter pages: 0 → 12
• Branded queries to the firm's name in ChatGPT (an indirect AEO signal): 6x increase, suggesting the firm's name was now part of ChatGPT's training/index recall for the vertical

Organic traffic was roughly flat (+4%), confirming the AI lift came from extraction-readiness, not from generic SEO improvements.

Why this worked: 5 transferable patterns

1. Direct-answer leads, every page. AI engines extract from the first ~80 words. A 312-word marketing intro guarantees you do not get cited. Lead with the answer, then expand.
2. FAQ schema where it earns its keep. Generic FAQ schema spamming does not work. FAQPage schema applied to genuinely useful Q&A produces citation lift; applied to filler, it can hurt.
3. Anonymized case studies as citable proof. LLMs reward specifics — jurisdiction, statute, outcome range. Fluffy testimonials are ignored.
4. Entity consistency across the open web. Directory NAP cleanup + sameAs array + Attorney schema with hasCredential make the firm a resolvable entity, not just a website.
5. Rank in the top 10 first, then optimize for extraction. Over 92% of AI Overview citations come from already-ranking domains. AEO without SEO is mostly futile.

Ethics and compliance notes (US-specific)

Law firms cannot copy-paste DTC AEO playbooks. Every rewritten page in this case study cleared a pre-publication ethics review against:

• ABA Model Rule 7.1 (no false or misleading statements)
• ABA Model Rule 7.2 (advertising and the comparison/superlative rules)
• State-bar specific advertising rules in every jurisdiction the firm practices

Key content rules the firm enforced:

• No outcome guarantees, even implicit (avoid phrases like "we win these cases")
• Settlement amounts expressed as ranges, not exact figures, and accompanied by the disclaimer required by the state bar
• Past results disclaimers placed adjacent to representative-matter blocks, not buried in the footer
• Attorney bios listed bar admissions explicitly (this also feeds the Attorney schema's hasCredential)

Firms in regulated verticals — legal, medical, financial — should treat AEO as a content + compliance program, not a content program alone.

What did not move the needle

• Generic AI-content-detection tools used to "AI-proof" copy. AI engines do not penalize AI-assisted writing; they penalize ungrounded, generic writing.
• HowTo schema on legal advice pages. Legal-advice content is rarely cited via HowTo — Google appears to suppress HowTo on YMYL legal topics. FAQPage and LegalService did the heavy lifting.
• Massive blog volume. The firm cut blog cadence from 12 posts/month to 4 deeper posts/month and saw AI citations rise. Volume without extractability is wasted spend.

Reproducible 90-day starter plan for a mid-size firm

• Days 1-14: Pick 20-40 pages already ranking 4-10 for AEO-friendly queries. Rebuild the citation baseline across AI Overviews, ChatGPT, and Perplexity using a 40-80 query panel.
• Days 15-45: Rewrite leads on every prioritized page to a 60-90 word direct-answer paragraph. Cluster subheads by long-tail phrasing.
• Days 30-60: Deploy FAQPage, LegalService, and Attorney schema. Validate with Schema.org validator and Google's Rich Results Test.
• Days 45-80: Publish 6-12 anonymized representative-matter case studies, ethics-reviewed.
• Days 60-90: Clean up directory NAP and sameAs. Earn 3-6 trade-press placements with deep links into practice-area pages, not the homepage.
• Day 90: Re-measure citation panel. Expect 1.5-2.5x lift on AI Overviews, first ChatGPT/Perplexity citations to start appearing on long-tail queries.

FAQ

Q: Is AEO different from SEO for law firms?

AEO targets citation and source-selection by AI answer engines, while SEO targets ranking on lists of blue links. In practice the two share foundations — top-10 organic ranking is a near-prerequisite for AI Overview citations — but AEO adds extraction-readiness, schema layering, and entity authority. Most law firms do not need to abandon SEO; they need to layer AEO on top.

Q: How long until a law firm sees AI citations after starting AEO?

In this case study the first new AI Overview citations appeared at week 6 and ChatGPT citations at week 11. Industry agency reports suggest 3-6 months is realistic for mid-size firms with existing top-10 rankings; 6-12 months for firms starting outside the top 10.

Q: What is the highest-ROI single change a law firm can make for AEO?

Rewriting the lead paragraph on every practice-area page to a 60-90 word direct answer to the page's primary question. This single change, applied across the priority page set, accounted for the majority of the citation lift in this case study.

Q: Does FAQ schema still work in 2026?

Yes for AI engines, even where Google has reduced FAQ rich results in classic SERPs. AI Overviews, ChatGPT, and Perplexity still parse FAQPage schema as discrete answer blocks, and pages with it are cited at higher rates than pages without it.

Q: Are AI citations worth it if users do not click?

For law firms, yes. AI citations function as branded recall: a prospective client who sees the firm cited in an AI answer is more likely to search the firm by name later. The case-study firm saw a 6x increase in branded ChatGPT queries in 5 months, even though direct click-throughs from AI surfaces were modest. Google has also made source links more visible inside AI Overviews, narrowing the click gap.