How to Check If ChatGPT Recommends Your Law Firm (And What to Do If It Doesn't)

# How to Check If ChatGPT Recommends Your Law Firm (And What to Do If It Doesn't)

This takes 15 minutes. Open ChatGPT, Perplexity, and Google. Run these seven prompts. Be honest about what you find.

If your firm appears in AI-generated responses for the buyer-intent queries your prospective clients actually type, you have AI visibility. If it doesn't — and for virtually every Towson-area law firm we've audited, it doesn't — you now have a baseline score of zero and a specific set of problems to fix.

The 7-Prompt Test

These seven prompts mirror the Magnificent 7 buyer-intent queries that real legal clients type into AI systems at the moment they need help. Run all seven on ChatGPT (chat.openai.com), Perplexity (perplexity.ai), and Google (look for the AI Overview box at the top of results). Substitute your actual practice area and location.

### Prompt 1: "Best [practice area] attorney in Towson MD"

What a citation looks like: AI names your firm specifically with a brief description of why — "Smith Criminal Defense has a strong presence in Baltimore County courts and focuses on first-offense DUI cases."

What mentioned looks like: Your firm's name appears in a list without specific recommendation.

What not cited looks like: AI names competitors, national resources, or gives a generic answer about how to find attorneys.

### Prompt 2: "What should I do after a [legal event] in Maryland"

For DUI: "What should I do after a DUI arrest in Maryland." For PI: "What should I do after a car accident in Maryland." For family law: "What should I do after being served divorce papers in Maryland."

What a citation looks like: AI references your content specifically — "According to [Firm Name], Maryland law requires you to request an MVA hearing within 10 days of arrest."

What not cited looks like: AI cites Nolo, FindLaw, Avvo editorial content, or other national sources with no local firm cited.

### Prompt 3: "How long do I have to [time-sensitive action] in Maryland"

For DUI: "How long do I have to request an MVA hearing after a DUI in Maryland." For PI: "How long do I have to file a personal injury lawsuit in Maryland." For family law: "How long do I have to respond to divorce papers in Maryland."

What a citation looks like: Your firm is named as the source of the specific deadline. The AI quotes your content by section number — "Under Maryland Transportation Article §16-205.1, you have 10 days."

What not cited looks like: Generic answer with no local source attribution.

### Prompt 4: "Top [practice area] law firms Baltimore County"

What a citation looks like: Your firm is listed with a specific reason for the recommendation.

What not cited looks like: A mix of Martindale-Hubbell ratings, Avvo editorial lists, and generic advice to "check reviews."

### Prompt 5: "Can a [event] affect my [career/license/status] in Maryland"

For DUI: "Can a DUI affect my teaching license in Maryland." "Can a DUI affect my nursing license in Maryland." "Can a DUI affect my CDL in Maryland."

What a citation looks like: AI attributes Maryland-specific information to your firm — citing PBJ, §21-902, or the professional licensing notification rules.

What not cited looks like: Generic explanation that doesn't mention Maryland-specific options like PBJ.

### Prompt 6: "Do I need a lawyer for [specific situation]"

"Do I need a lawyer for a first-offense DUI in Maryland." "Do I need a lawyer to respond to divorce papers in Maryland."

What a citation looks like: Your firm is cited as the source of the answer — ideally a specific, quoted statistic about outcomes with vs. without legal representation.

What not cited looks like: Generic pro-lawyer argument from a national source.

### Prompt 7: "[Firm name] reviews"

What a citation looks like: AI summarizes your firm's reputation from verified review sources with specific quoted feedback.

What not cited looks like: AI says it can't find reliable information or returns nothing.

How to Run Each Prompt

On ChatGPT: Go to chat.openai.com. Start a new conversation. Type each prompt exactly as written, substituting your practice area and location. Note whether your firm is named, mentioned, or absent. If you have a ChatGPT Plus subscription, also run the prompts with Browse enabled to test the web-retrieval version.

On Perplexity: Go to perplexity.ai. Type each prompt. Perplexity shows its sources — look at the source citations in the right panel. Is any source from your firm's domain? If the sources listed are all nolo.com, justia.com, or findlaw.com, your content is not being retrieved.

On Google: Search each prompt. Look for the AI Overview box that appears above the organic results. Note whether your firm is named in the AI Overview and whether it appears as a cited source with a small icon and link. Then scroll to the organic results — your position there is a separate metric from your AI Overview citation.

Document everything. Make a simple table: Prompt | ChatGPT | Perplexity | Google AI. Mark each cell: Cited (named as source), Mentioned (named without attribution), or Not Cited.

How to Score Your Results

Count the total number of Cited results across all 7 prompts and all 3 platforms (maximum 21 possible citations).

17-21 citations: AI-DOMINANT. Your firm is being recommended across the leading AI platforms for all major buyer-intent queries. You have strong MAKIF-46 signals and a significant competitive advantage. Focus on Measurement — tracking and defending your position as competitors wake up.

10-16 citations: CITATION-ELIGIBLE. Your firm is being cited for some queries on some platforms. You have meaningful AI visibility but gaps. Run the MAKIF-46 Audit to identify which specific signals are causing the gaps.

4-9 citations: STRUCTURED. Your firm is mentioned occasionally but rarely cited as the authoritative source. You have some foundational signals in place but significant content and entity gaps. A 60-day focused program can move you to Citation-Eligible.

0-3 citations: INVISIBLE. This is where virtually every Towson law firm we've audited sits. You are not being cited by any major AI platform for any of your practice area's highest-intent queries. Your MAKIF-46 score is likely 7-20. You need foundational work across Entity, Retrieval, and Citation layers before any other marketing investment compounds.

What to Do If You're Not Cited

Fix in this diagnostic order:

Step 1: Entity. Before content matters, AI systems need to know who you are. Check your Google Business Profile (claimed and verified?), your schema markup type (LegalService, not LocalBusiness?), and your NAP consistency across Avvo, Justia, FindLaw, and Maryland State Bar. Fix these first. They take 1-2 weeks and establish the foundation.

Step 2: Content Structure. Audit your practice area pages. Do any of them answer a specific legal question in the first paragraph? Do any cite a Maryland statute by section number? Do any have FAQ schema? If no to all three, your content is not retrieval-ready. The fix: rewrite or supplement one page per practice area to the direct-answer format, with FAQ schema and at least one statute citation.

Step 3: Citation Triggers. Add specific, sourced data points to your content. For DUI: "Maryland recorded 23,000+ DUI arrests in 2024, according to NHTSA and Maryland MVA data." <cite><a href="https://www.nhtsa.gov/research-data/drunk-driving" rel="external">NHTSA/Maryland MVA, 2024</a></cite> For PI: "Maryland Courts and Judicial Proceedings §5-101 establishes a 3-year statute of limitations for personal injury claims." These specifics are what AI systems quote. Generic statements are what they skip.

Step 4: Measurement. Set a recurring calendar reminder to run the 7-prompt test every Monday. Document results. This is how you know whether your fixes are working.

Why Your Competitors Are Getting Cited

In almost every Towson law firm market segment we've examined, the firms getting cited by AI are not local firms at all. They are national legal information platforms — Nolo, FindLaw's editorial guides, Justia's legal explainer content — that publish direct-answer content with citations at scale.

The reason these platforms win: they publish content that is explicitly structured for retrieval. Direct answers in the first paragraph. Statute citations by section number. FAQ schema. Named authors. Source citations. Content updated regularly.

None of this requires being a national platform. A single Towson DUI firm that publishes one correctly structured page on "What Is Probation Before Judgment in Maryland?" — with a BLUF answer, §21-902 citation, PBJ definition, FAQ schema, and named attorney attribution — will compete directly with Nolo for that specific query. And it will win for the local context, because AI systems prefer geographically specific authoritative content for local intent queries.

The entity + FAQ schema combination is the core of what's working. Correct schema type tells AI what you are. FAQ schema tells AI where your direct answers are. Together they create the retrieval pathway that national platforms have and local firms lack.

Frequently Asked Questions

*How often should I run the 7-prompt test?*

Weekly, at minimum. AI citations are not static — they shift as new content is published, as AI systems are updated, and as competitors begin optimizing. A firm that runs the test monthly may miss a three-week window where a competitor gained citation share. Set a Monday morning calendar reminder and document results in a simple spreadsheet.

*What does it mean if Perplexity shows my domain as a source but ChatGPT doesn't?*

It means your content is being retrieved by Perplexity's crawlers but not reflected in ChatGPT's training data or retrieval layer. This is common for recently published or recently updated content — Perplexity uses live web retrieval while ChatGPT's base model has a training cutoff. Fix: ensure your content is indexed in Google Search Console, then submit for crawling. ChatGPT's Bing-powered retrieval should pick it up within 2-4 weeks.

*Is it bad if AI gives my firm's name without attribution?*

Mentioned without attribution is better than invisible, but it's not a citation. An AI that says "You might also look at Smith Criminal Defense in Towson" is not citing your content as authoritative — it's surfacing your name from a directory. Citations — where AI references your specific content as the source of a legal fact — are the meaningful metric for GEO. Work to convert mentions to citations through content and FAQ schema improvements.

*What if I run these prompts and a competitor shows up every time?*

That competitor has done some GEO work — or they have content signals that happen to be AI-retrievable. Note which queries they're winning and examine their content on those topics. Are they citing statutes by section number? Do they have FAQ schema? Is there a named attorney on the page? That's your competitive intelligence. Match and exceed those signals on your own pages.

*How do I know if my Google Business Profile is affecting my AI visibility?*

Test it directly: search your firm name in Google and look for a Knowledge Panel (the box on the right with your hours, address, and photos). No Knowledge Panel means weak entity recognition. Also check whether your Business Profile categories correctly reflect your practice areas. A family law firm categorized only as "Legal Services" is less entity-clear than one categorized as "Family Law Attorney" with individual attorney profiles linked.

*Does the 7-prompt test work for all practice areas?*

Yes, with adjustments. For each practice area, substitute the relevant legal event (DUI arrest, car accident, divorce service, estate planning need) and the relevant time-sensitive action (MVA hearing, PI lawsuit filing, divorce response, Medicaid application). The structure of the test is the same; the specific prompts are practice-area specific.

*What is the difference between ChatGPT Browse mode and standard mode?*

Standard ChatGPT uses its training data (cutoff: early 2024 for most models). Browse mode uses live web retrieval through Bing. For testing AI visibility, Browse mode is more relevant because it reflects how AI will answer questions using current web content. Run prompts in both modes — the gap between your standard mode citation rate and your Browse mode citation rate tells you how recently your content has been indexed and whether your retrieval signals are working.

*What should I do immediately after running the test and scoring zero?*

In order: (1) File an llms.txt at your domain root — this takes 10 minutes and signals to AI crawlers what content is available. (2) Fix your schema type to LegalService. (3) Submit your existing practice area pages to Google Search Console for immediate crawling. (4) Book the MAKIF-46 Audit for a full diagnostic. These four steps can be completed in one afternoon and address the most acute failure modes.

Want a full diagnosis beyond the 7-prompt test? [Book the MAKIF-46 Audit](/audit) for a scored assessment across all 46 signals.

Sources: KDD '24 GEO Research (Aggarwal et al., arxiv.org/abs/2406.13692) · SE Ranking AI Overviews Research 2024 · WebFX / Semrush AI Referral Traffic Research · NHTSA/Maryland MVA 2024 · MAKIF Audit Data, Towson MD, 2025-2026