Last updated July 2026
When a general counsel types “best restructuring firm for a $300M mid-market deal in Texas” into ChatGPT, your firm’s website is almost certainly not what gets cited. Chambers is. Legal 500 is. A Law360 profile piece is. Your bio page on the firm website is not.
That is the citation gap that practice-area AEO fixes.
This playbook walks you through four steps: running the right prompt set, mapping where citations are going, building content engineered for AI extraction, and earning your way into the third-party sources AI engines trust.
Step 1: Run the buyer-intent prompt set
Most law firms track keyword rankings for terms like “M&A attorney Chicago.” That is not how general counsel, procurement teams, or boards phrase a search in ChatGPT.
They ask:
- “Which law firm handles cross-border energy M&A deals for mid-market companies?”
- “Best IP litigation firm for semiconductor patents in the Northern District of California?”
- “Top restructuring counsel for distressed real estate in New York?”
These are the prompts you need to run, repeatedly, across multiple AI engines. Each one is a buyer-intent query with a specific practice area, deal type, and jurisdiction. They are the prompts that produce citations.
Temso is the practical starting point here. At $89/mo, it tracks a custom prompt set across eight AI engines (ChatGPT, Perplexity, Google AI Overviews, Google AI Mode, Gemini, Grok, Microsoft Copilot, and Meta AI), runs each prompt multiple times to stabilize the citation rate, and shows you which domains are being cited instead of yours. For a law firm marketing team that does not have a dedicated AEO specialist, it covers the full monitoring loop without requiring technical setup.
Peec AI is the stronger choice for multi-jurisdiction or multi-practice-area programs. Its citation-gap queue converts monitoring results into a prioritized task list: for each prompt where a competitor or directory is being cited instead of you, it shows you what to do next.
Profound is built for enterprise legal teams with AEO-dedicated headcount. Its Prompt Volumes feature shows what millions of users are actually asking AI systems, so you can build your prompt set from real demand data rather than guessing.
Build the prompt matrix
Organize your prompts into three layers:
| Layer | Example | What it tells you |
|---|---|---|
| Category | ”Best M&A law firm in [city]“ | Baseline citation share for the practice |
| Qualifier | ”Top restructuring firm for mid-market deals” | Whether deal-type specificity helps or hurts your citations |
| Situation | ”Which firm handles SPAC litigation for tech companies?” | Whether you win on specific scenario queries |
Run each prompt at least five times per engine. Citation rates are probabilistic. A single run is noise, not signal. For more on why, see how AI citations work.
Step 2: Map the citation gap
Once you have run the prompt set, you will see a consistent pattern. The same handful of domains win the citations for legal queries, and almost none of them are law firm websites.
The pattern holds across practice areas: ranking directories dominate (Chambers, Legal 500, Super Lawyers, Best Lawyers, Martindale-Hubbell, Avvo, Justia), followed by legal news outlets (Law360, Above the Law, The American Lawyer), and then bar association resources. Firm-owned websites rarely appear for “best firm” queries regardless of how well-written or keyword-optimized they are.
Your citation-gap map should answer three questions:
- Which third-party domains are cited for your target practice and jurisdiction?
- Are competitors cited on those domains? How prominently?
- Where are you cited today, and for which qualifiers?
This map becomes your editorial calendar for earned-media outreach. Every domain that wins citations for your target queries is a publication you should be contributing to, being profiled in, or earning a directory listing on.
Step 3: Publish content engineered for AI extraction
Law firm websites are written by lawyers for lawyers. That produces dense, qualified, credential-heavy content that is hard for AI systems to extract a direct answer from.
Jurisdictional guides are the format that fixes this. A jurisdictional guide answers one bounded question: “How does [practice area] work in [state or jurisdiction]?” It is the format AI engines cite because the extraction job is easy: the first paragraph is the direct answer, the body is the process, and the FAQ at the bottom is already in question-and-answer form.
The jurisdictional guide structure
A well-engineered jurisdictional guide has these sections, in this order:
-
Direct answer (40 to 60 words). State the core answer to the title question in the opening paragraph. This is what AI engines quote. Do not bury it.
-
Deal-parameter context. Specify the transaction sizes, timelines, and local rules that apply. Specificity is the signal. “Typical M&A deals in Texas with an enterprise value of $50M to $500M close in 60 to 120 days under the TBCA” is extractable. “Our experienced team handles a range of transactions” is not.
-
Process outline. Walk through the stages of the matter in numbered steps. Each step should be one sentence answering what happens and why it matters. Numbered structures are easy for AI systems to pull into a summarized answer.
-
FAQ block with schema. Write four to eight questions that a buyer would actually ask at this stage of evaluating a firm. Each answer should be two to four sentences: direct, specific, and free of hedging. Add FAQPage JSON-LD schema to signal to AI crawlers that the page is structured as question-and-answer pairs.
-
Outcome data. If your firm can share deal parameters (size range, jurisdiction, industry), do it. Anonymized deal data is credible and extractable. “Represented a private equity sponsor in the $140M acquisition of a Texas-based SaaS company, including negotiation of rep-and-warranty insurance and a custom earnout structure” gives AI systems something specific to cite.
The extraction-first writing rule
According to Kevin Indig’s February 2026 analysis of 1.2 million ChatGPT responses, 44.2% of ChatGPT citations were drawn from the first 30% of a page’s content. Put your most citable sentences at the top of every section, not at the end.
Every H2 should complete the sentence “The answer to this heading is…” before expanding. Every paragraph should have its point in the first sentence. Write for extraction, then for reading.
FAQ schema: worth doing, not a magic fix
Add FAQPage JSON-LD to your jurisdictional guide and your practice-area pages. It helps AI crawlers parse the content correctly. It does not guarantee citation uplift by itself. An Ahrefs controlled study of 1,885 pages (2025 to 2026) found no statistically significant increase in AI citations from adding schema alone. Schema is hygiene, not a shortcut.
For tools that generate and validate FAQPage schema as part of a workflow, Temso includes this in its base plan.
Step 4: Earn citations through third-party outlets
Publishing a well-structured jurisdictional guide on your firm website is necessary but not sufficient. AI engines cite third-party sources for “best firm” queries. You need to appear in those sources.
A December 2025 Stacker and Scrunch pilot study found that distributing content through third-party news outlets lifted AI citation rates from roughly 8% (for brand-owned content) to 34%, a 325% increase. A larger Stacker follow-up in March 2026 confirmed the direction of the effect at a 239% median lift across a broader dataset. The study covered general earned media across 17 industries, not legal specifically, but the mechanism is the same: AI engines encounter your content in publications they already trust and cite, so your content inherits that trust.
For law firms, the third-party amplification stack looks like this:
Tier 1: ranking directories. Chambers, Legal 500, Super Lawyers, Best Lawyers. These are the sources AI engines cite most for “best firm” queries. Nominations, submissions, and client referrals into these directories are not just credentialing exercises. They are citation-building infrastructure.
Tier 2: legal trade publications. Law360, Above the Law, The American Lawyer, and regional legal news outlets. Bylined articles, contributed commentary on recent cases, and deal announcements in these outlets create AI-citable content that links the firm’s name to specific practice areas and jurisdictions.
Tier 3: bar and industry publications. State bar journals, practice-area industry publications (for the firm’s client sectors), and association resources. These carry authority signals that AI engines recognize.
Tier 4: structured third-party profiles. Martindale-Hubbell, Avvo, and Justia profiles should be complete, current, and consistent with the practice-area language on your own site. Entity consistency across sources reinforces the signal that your firm is a recognized authority in the named practice areas.
Turning earned media into a system
Most law firm PR and business development efforts produce scattered placements. AEO requires a systematic approach: track which publications produce citations for your target prompts, prioritize outreach to those publications, and publish content in a format those publications can use.
Contributed articles that include deal-parameter data, jurisdictional context, and a clear process explanation produce the most citable third-party content. A 1,200-word bylined piece in Law360 titled “How Texas Courts Are Handling Earnout Disputes in Post-Acquisition Litigation” gives both the publication and AI engines something specific to surface.
Scrunch AI is useful at this stage for enterprise legal teams. Its Agent Experience Platform serves AI-optimized content directly to LLM crawlers, bridging the gap between what you publish and what AI systems actually read. For firms managing significant volumes of practice-area content, it handles the technical delivery layer.
Putting it together: the practice-area citation loop
These four steps form a loop, not a one-time project:
-
Run the prompt set monthly. Citation patterns shift as AI engines update, as competitors publish, and as new directories or publications earn authority.
-
Review the citation gap. Add new domains to your earned-media target list. Remove or deprioritize ones that no longer win citations.
-
Publish or update the jurisdictional guide. Treat it as a living document. Add new deal-parameter data as matters close. Update the FAQ as client questions evolve. AI systems weight fresh content.
-
Execute third-party placements. Track which placements produce citations by re-running the prompt set after each piece publishes. This is how you measure earned-media ROI in an AEO program.
The full loop is covered at /methodology, and the AEO tools ranking has a current comparison of the monitoring platforms that fit different firm sizes.
Start with the prompt set. Run it this week. The citation gap will tell you exactly where to spend the next 90 days.