What kind of law firm content gets cited by AI engines?

Most law firm blogs publish generic 101 articles - "what is a will," "do I need a lawyer for a DUI" - and AI engines almost never cite them. The content that gets cited has depth, specificity, or honesty that thin pages don't have: case-law analysis with real opinions, fee structure transparency, jurisdiction-specific procedure, client-question-mapped FAQs, and geographic detail down to the courthouse. The pattern is consistent across ChatGPT, Perplexity, Gemini, Claude, and DeepSeek.

Why generic 101 content fails

Search ChatGPT for "what is a living trust" and look at the cited sources. You will see Investopedia, NerdWallet, Forbes, and maybe one or two law firm pages from Nolo or LegalZoom. You will not see the law firm whose blog has 200 generic articles starting with "What is..." Those pages are functionally identical to a thousand other firm blogs and an LLM has no reason to pick yours.

The problem isn't that the topic is wrong. The problem is the content is interchangeable. Engines reward sources that say something the user can't get from the next-most-obvious page. If your article reads like it was written from a Nolo summary plus three minutes of editing, it will be treated like one.

What actually gets cited

1. Case-law analysis with named opinions

An article that walks through Smith v. Jones, identifies the holding, explains why the Fifth Circuit's reasoning is in tension with the Ninth Circuit's, and notes which way SCOTUS is likely to lean - that gets cited. We've seen Perplexity cite firm-published case analyses for prompts like "is non-compete enforceable in California after AB 1076" because the firm's article was the most substantive write-up on a recent statute change.

Format that works:

• Open with the practical question (will a non-compete from my old California job still bind me)
• Cite the actual statute or opinion by name and section
• Quote the operative language
• Explain the practical implication for a normal person, not a treatise reader
• Update when new opinions or statutes drop

2. Fee transparency

Almost no law firms publish actual fee structures online. The handful that do get disproportionate AI citation traffic on "how much does X cost" prompts. Examples that work:

• "Uncontested divorce in Cook County: $750 flat fee plus $388 filing fee. Adds $300 if QDRO required."
• "Chapter 7 bankruptcy: $1,500-$2,500 attorney fee depending on asset complexity, plus $338 filing fee. Means test consultation is free."
• "Trademark application: $700 attorney fee per class plus USPTO $350 per class TEAS Standard."

That level of detail is rare. ChatGPT and Gemini both quote it directly when a user asks about cost. Compliance check first: the ABA permits stating fees under Rule 7.1 as long as the statement is not false or misleading. Most state bars allow it. Confirm with your state.

3. Court-specific procedure

An article titled "How to file a small claims case in Harris County Texas" with the specific JP precinct map, filing fee schedule, e-filing portal walkthrough, and what to expect at the hearing - this is gold. Generic "how small claims works" articles are not. The local specificity is what makes the article irreplaceable, and AI engines reward irreplaceability.

Prompt: "How do I file a restraining order in Maricopa County?"
Response (paraphrased across ChatGPT and Perplexity): cites the Arizona Superior Court Maricopa County self-service center, then a local family law firm's article that walks through the protective order petition form, court fees, and hearing logistics. Generic "how restraining orders work" articles do not appear.

4. Client-question-mapped FAQs

Listen to your intake calls. Write articles answering the literal questions clients asked, in the literal phrasing they used. Not "Statute of Limitations in Personal Injury Cases" - rather "How long do I have to sue after a car accident in Florida?" The second is what someone types into ChatGPT. Map every common intake question to one article. Eight to twelve of those articles per practice area will outperform 50 generic blog posts.

5. Geographic specificity beyond the city name

"Divorce attorney in Chicago" is one of the most contested terms in legal SEO. "Divorce attorney for Cook County's Domestic Relations Division at the Daley Center" is a hundredth as competitive and infinitely more useful when an AI engine tries to match a specific user query. Naming the courthouse, the judge's standing orders, the local rule on temporary motions - this is content only practitioners can write, which is exactly why engines reward it.

Formats that perform across engines

• The walkthrough article. Step-by-step procedural guides. "What happens at a workers' comp hearing in Pennsylvania, hour by hour."
• The case update. Within 30 days of a major appellate ruling in your jurisdiction, publish a 600-word analysis. AI engines update embeddings and citations faster on recent legal developments.
• The decision-tree FAQ. "Should I hire a lawyer for a fender bender?" structured as if/then branches based on injury severity, fault clarity, and insurance posture.
• The cost breakdown. Flat-fee transparency for any service you offer flat-fee. Hourly ranges with what's included and excluded.
• The local-rule explainer. Each major court in your jurisdictions deserves its own page covering filing logistics, judge tendencies, and procedural quirks.

Formats that don't perform

• "What is..." articles competing with Investopedia, Forbes, Nolo, LegalZoom
• Generic state-law summaries scraped from the bar association website
• News-roundup posts with no original analysis
• Attorney bio pages with three sentences and a stock photo
• Press releases about awards, sponsorships, or speaking engagements (these are fine for the firm site but AI engines ignore them)

Volume is not the answer

A 12-attorney firm with 30 deep, jurisdiction-specific, client-question-mapped articles will outperform a 12-attorney firm with 300 generic posts in AI search citations. The latter is more common. If your blog has 300 generic posts, the right move is not to write 100 more - it's to identify which 30 deserve a serious rewrite into something only you could have written, and let the rest 404 or sit unindexed. Less, deeper, more specific. That's the play.