Prerequisites
Before tracking a single citation for chatgpt visibility, the firm had the non-negotiable pieces in place. Their existing SEO program had been running for four years. Domain authority sat at 53. They held Martindale-Hubbell AV ratings, maintained a deep blog with 180+ fact-checked pages, and all attorneys held active state bar memberships. The site was mobile-optimised, loaded in under 2.1 seconds on 4G, and carried no manual actions. That foundation isn’t optional when entity recognition drives AI engine citations.
The only gap: zero presence in AI-generated answers. ChatGPT, Perplexity, Gemini, and Claude either ignored the firm or cited three competing practices for queries like “best personal injury lawyer in [city]” and “what to do after a car accident.” Traffic from organic search was stable, but zero-click queries were rising. A look at the death of the blue link made the urgency tangible.
What they didn’t need A new website. No full content rewrite. No paid ads. The assets that rank on Google already map to what AI engines extract, they just weren’t structured for machine answer generation.
Step 1: Audit chatgpt Visibility with Engine-Specific Data
Read the room before rewriting a word. The legal marketers pulled a 90-day baseline across ChatGPT, Perplexity, Gemini, and Claude. They used the AEO Scorecard’s 12 metrics to measure:
- Citation share: how often their firm name appeared in AI answers vs. competitors for 50 seed queries.
- Engine coverage: which engines cited the firm (initially, only Perplexity on 2 queries).
- SERP-to-AI displacement: for which high-volume queries was the blue link pushed below the fold by an AI-generated answer.
Three findings forced the strategy:
- Competitors with thinner content got cited because they used explicit Q&A formatting and short answer paragraphs at the top of pages.
- ChatGPT overwhelmingly relied on a tight set of authoritative directories plus a handful of legal-specific sites like Avvo and Martindale-Hubbell, neither of which the firm had optimized for AI extractability.
- 34% of the firm’s top organic queries now appeared with an AI answer box that didn’t mention them.
Pro tip
Run the same audit at the query level, not just the domain level. A firm might get cited for “car accident lawyer” in Perplexity but not in ChatGPT, and that gap tells you where your entity signals are weak. Tools like SearchFit’s “Brand Presence” report break this down by engine.
Warning
Don’t conflate “AI visibility” with “brand mentions in AI-generated text.” An AI can describe your firm without citing it as a source. Only measure instances where the model attributes content or names the firm as a recommendation. The distinction matters when you measure AEO ROI for your CFO.
Step 2: Map the Questions Real Clients Ask AI
Law firm SEO typically targets “near me” and “best” keywords. AI engines answer phrasing differently. The team executed prompt research, mining real questions users ask ChatGPT, Perplexity, and Gemini in their practice area. They collected prompts from:
- AI conversation logs shared by two intake coordinators who asked clients “Did you use any AI tool before calling us?”
- Public Reddit and Quora threads where users described queries they tried on ChatGPT legal topics.
- SearchFit’s “Prompt Explorer” dashboard, which surfaces high-frequency AI queries by vertical (personal injury, family law, estate planning).
They identified 147 distinct prompts. Examples:
- “What’s the average settlement for a rear-end collision in California?”
- “Can I sue if I was partly at fault in a motorcycle crash?”
- “How long does a personal injury lawsuit take in Texas?”
These didn’t match the firm’s existing keyword list. Google targeting produced “car accident lawyer Houston,” not “process to file a lawsuit in Texas.” The gap was clear. The team prioritized 30 high-volume AI queries and designed content modeled after the exact prompt structure. The LLM SEO for Law Firms guidance on SEJ explained that natural language, scenario-based content beats keyword-optimized pages for AI retrieval. They applied it.
Step 3: Build Extractable, AI-Ready Content Hubs
Each priority query became a hub page following a strict template:
- Answer-first paragraph: 50-60 words answering the query directly, with a citation to a primary legal source. For the question on fault, the snippet reads: “Texas follows modified comparative negligence with a 51% bar. You can sue for damages as long as you are 50% or less at fault. (Source: Texas Civil Practice & Remedies Code § 33.001).”
- Expanded scenario blocks: Realistic fact patterns with attorney commentary, citing actual case outcomes.
- FAQ section with numbered answers: each question formatted as an H3 with a 3-4 sentence response, avoiding fluff.
- Schema: FAQPage structured data on every hub, plus LegalService schema tying each attorney to the firm entity.
The firm ran this structure through Google’s helpful content guidelines and through the lens of AI hallucinations prevention. Because AI models privilege demonstrable accuracy, every sentence that stated a dollar figure, statute, or procedural deadline was linked to an authoritative government (.gov) or bar association source. This isn’t just risk management; it’s an intentional signal to language models that the content is trustworthy. As the ABA notes in its AI and legal practice coverage, factual precision is the baseline for ethical legal marketing in the AI age.
Warning
Never publish AI-generated legal advice, even for topical authority. Any content drafted by a language model must receive attorney review and be clearly labeled where required. The California Privacy Rights Act also mandates transparency about automated decision-making if it impinges on client rights. The firm’s operations manager enforced a “reviewed by [attorney name]” footer on every page and ran quarterly audits with the Legal Ethics Institute framework.
Step 4: Cement Entity Authority Inside AI Training Data
LLMs build their internal knowledge graphs from high-trust domains. For legal, that means directories, bar associations, and academic publications. The firm executed five moves:
- Sync NAP and attorney profiles on Martindale-Hubbell, Avvo, Justia, and FindLaw. They made sure each profile linked to the firm’s site and used identical attorney names, descriptions, and practice area tags. This directly feeds the Knowledge Graph that Google and ChatGPT ingest.
- Claim and enhance the Google Business Profile with weekly posts answering common legal questions. Those posts appear in Gemini’s local knowledge panel extractions.
- Publish three articles on the ABA’s Business Law Today digital platform, each with a canonical link back to the firm’s expanded hub page. The ABA’s domain authority is a top input signal for entity SEO.
- Submit amicus brief outlines on state bar ethics opinions to the NIST AI Risk Management Framework comment repository, citing the firm’s role in shaping responsible AI in legal. (Yes, it’s niche. It generated two ChatGPT citations the following quarter.)
- Build deep, not wide. Instead of 1,000 thin pages, they published 47 long-form hubs with 2,500+ words each, all interlinked by topic. Internal linking passed entity strength, signaling to crawlers that this firm holds authoritative knowledge on specific legal scenarios.
After 60 days, AI citation tracking showed a measurable shift. The firm appeared in 8 ChatGPT answers per week, up from zero. Perplexity citations doubled. Gemini started citing two of the firm’s articles in legal “additional reading” snippets.
Step 5: Monitor, Defend, and Recover Citations Weekly
AI citations aren’t static. A model update can drop a firm overnight. The team established a Friday ritual:
- Pull the AEO Scorecard for the top 30 tracked queries. Note any engine where the firm lost a citation.
- Check competitor citation movements. If a competitor replaced the firm, reverse-engineer their page structure. Was it a fresher date? A more succinct answer? A new directory profile?
- Update the rapid-response content artifact. For any page that lost a citation, add a new expert quote, update the publication date, and resubmit the URL to Google’s API. Semantic freshness signals boost AI retrieval within 10-14 days.
- Revalidate schema markup. A dropped citation often traces back to a broken Schema node or a site structure change that broke entity linking.
When a Claude update in late Q3 removed the firm from 11 answers, the team used the loss recovery playbook to regain 9 of those within 13 days. The fix: condensed the top-answer paragraph to 47 words (from 82), added a direct statute citation, and updated the mentioned settlement range with Q3 court data. Accuracy plus brevity won.
Pro tip
Set up real-time alerts for brand name + “according to” mentions across AI output. If your firm name appears but not as a source, that’s a warm lead: publish a dedicated page matching the query and request recrawling.
Step 6: Scale with Dashboards and API-Driven Visibility
Once the manual playbook proved ROI, the firm’s marketing ops lead automated visibility tracking using SearchFit’s API. They built internal dashboards, modeled after this agency’s white-label AEO dashboard, to show:
- Citation trendlines by engine in a single pane for the managing partners.
- AI-driven lead attribution: intake source “asked ChatGPT” grew from 0% to 12% of total calls after six months.
- Competitive citation gap: side-by-side engine share for the firm vs. top 3 competitors.
This shifted the conversation from “is AEO real?” to “how do we double citations next quarter?” The firm’s CFO now reviews AEO KPIs alongside paid search CAC. The numbers: post-implementation monthly qualified leads from organic channels increased 40%, and 63% of new clients reported “found you when I asked [AI tool] for a lawyer.”
Results and Key Takeaways
In seven months, the firm moved from invisible to cited weekly across every major AI answer engine. Specifics:
- ChatGPT citations: 0 → 12/week for priority queries.
- Perplexity citations: 2 → 19/week.
- Gemini local citations: appeared in “best personal injury lawyer” packs for 3 metro areas.
- AI-attributed leads: 12% of intake, with a 22% close rate, 8 points higher than paid search.
What made the difference
- Treating AI engines as independent discovery channels, not side effects of Google SEO.
- Building extractable, structured answers backed by primary legal sources.
- Continuously monitoring and defending citations, just like you would organic rankings.
- Leveraging entity authority through directories, bar associations, and third-party platforms like Avvo and Martindale-Hubbell.
The next 12 months
The firm is now optimizing for voice-based legal queries on the same AI engines and testing content strategies for Grok’s new real-time web integration. The playbook that earned the first 12 citations scales. The complete AEO guide shows the blueprint. For law firms, AI visibility isn’t a future risk, it’s a current reality where the firms that show up first in ChatGPT’s answer box take the privilege of being the only name the client remembers.