A law firm GEO case study should end in signed clients, and this one does. A Baltimore personal injury firm had 5 AI leads and 0 signed. It tracked 2 signed clients to ChatGPT, with +425% AI visibility growth and 4 mo. time to attribution. Here is the play, lever by lever, and how WKND executes it.

INFORM: ChatGPT sends you leads, and none of them sign

You pay for SEO or ads. ChatGPT is now sending some of your leads. If they are the wrong case types from the wrong places, you pay for a phone that rings and a calendar that stays empty.

That was the Baltimore firm. It handles car accidents, slip and fall, medical malpractice and wrongful death, with a strong local reputation. When people asked ChatGPT or Perplexity who handles serious injury cases in Baltimore, the firm did not appear. When it did appear, it was for jurisdictions and case types it doesn’t serve. Five AI leads arrived. None signed.

Its overall AI visibility across Baltimore personal injury prompts sat at 4%. It had zero Spanish-language presence on any AI engine. Leads were logged in a spreadsheet, source unknown.

The one number that matters: 2 signed clients attributed to ChatGPT.

Key figures

  • +425% AI visibility growth, sustained across all four AI engines.
  • 56% visibility on the top Baltimore slip and fall prompt, with 100% Share of Voice.
  • 2 signed clients attributed to ChatGPT, tracked in LeadDocket.
  • 4 mo. time to attribution.
  • Before: 5 AI leads, 0 signed, 4% overall AI visibility across Baltimore personal injury prompts.

EDUCATE: the four levers behind 2 signed clients

LeverThe problem it fixedWhat changed
Real prompts4% overall AI visibilityTargeted the specific prompts Baltimore injury victims use
Right market5 AI leads, 0 signed: wrong case types and jurisdictionsVisibility built for Baltimore and the case types the firm handles
LanguageZero Spanish-language presence on any AI engineBilingual authority built with original Spanish-language signals
AttributionLeads logged in a spreadsheet, source unknownEvery signed client tracked from first AI touchpoint through intake into LeadDocket

Lever 1: Target the prompts people actually type

Hook: Being visible is not the goal. Being named for the right question is.
Problem: General visibility fills your intake queue with people you can’t serve.
Solution: Pick the specific prompts a Baltimore injury victim asks, and build answer authority for each one.
Proof: The firm ran a five-pillar AI visibility system built for personal injury firms, aimed at those prompts. It reached 56% visibility on the top Baltimore slip and fall prompt with 100% Share of Voice. Best personal injury attorney in Baltimore reached 46%. Who handles truck accident cases in Baltimore reached 38%. Medical malpractice lawyer, Baltimore, Maryland reached 28%.

Lever 2: Aim at the cases and jurisdiction you serve

Hook: A lead you can’t take is a cost.
Problem: The first 5 AI leads were the wrong case types and jurisdictions. 0 signed.
Solution: Build visibility around your city and the case types you handle: car accidents, slip and fall, medical malpractice and wrongful death.
Proof: The firm’s tracked prompts now name Baltimore and its case types: slip and fall at 56%, truck accidents at 38%, medical malpractice at 28%. Result: 2 signed clients attributed to ChatGPT.

Lever 3: Win the language no one else is answering

Hook: The biggest gap in your market may be a language.
Problem: Spanish-speaking injury victims in Baltimore had no AI-visible firm with intentional Spanish authority, in one of the Mid-Atlantic’s largest Spanish-speaking markets.
Solution: Build original Spanish-language signals so your firm surfaces when a Spanish speaker asks AI for help after an accident.
Proof: The Spanish-language “best personal injury lawyer in Baltimore” prompt reached 44% visibility, from zero.

Lever 4: Write down where every client came from

Hook: If you can’t see the source, you can’t count the win.
Problem: With leads in a spreadsheet and the source unknown, a ChatGPT client looks like a referral.
Solution: Track every signed client from the first AI touchpoint through intake and into your case system.
Proof: “Every signed client’s origin is now tracked from first AI touchpoint through intake and into LeadDocket.” That is how 2 signed clients became 2 signed clients attributed to ChatGPT, with 4 mo. time to attribution.

How WKND executes this play

This is how WKND executes it. WKND builds exactly this system for law firms inside its legal SEO program. The AI SEO for Law Firms in Sacramento page says it this way: “Underneath it sits the same legal SEO program that ranks your case pages.”

Step by step, with WKND’s capabilities verbatim:

StepWhat WKND builds
1. Answer audit“Identify every featured snippet, PAA box, and AI Overview in your target keyword set”
2. Opportunity mapping“Prioritize answer opportunities by search volume, competitive difficulty, and business value.”
3. Content restructuringPart of the six-step program
4. Schema deploymentPart of the six-step program
5. New answer content“Create new content specifically designed to capture unclaimed featured snippets and AI answer positions.”
6. Monitor and defend“Track snippet ownership, detect losses, and rapidly respond with optimized content to reclaim positions.”
Intake attribution“Form and call tracking that reports cost per signed case instead of raw traffic counts.”

On top of that, WKND builds “AI Overview optimization. Optimize content structure and authority signals to be cited in Google’s AI Overviews and other AI-generated answer boxes.” For presence inside AI models, WKND’s LLM optimization service delivers “Brand presence engineering across ChatGPT, Gemini, Claude” and “Sentiment and recommendation monitoring across AI models.” The answer engine optimization service covers snippets and People Also Ask.

Set against the four levers, WKND runs the same loop: audit the prompts, map the ones that carry cases, build the answer content, and track every client to a source.

How do you track AI visibility and signed clients?

Track both, in two columns:

  • Google Analytics. Google’s release notes show a dedicated AI Assistant channel added to the Default Channel Group on May 13, 2026.
  • Search Console. Google’s generative AI report shows impressions of your links in AI Overviews and AI Mode, rolled out to all websites worldwide on August 31, 2026.
  • Intake. Ask every caller and form how they found you, and log “ChatGPT” as an answer.
  • A fixed prompt panel. Run the same real client questions every month and log whether your firm is named.

Visibility leads. Signed cases pay.

INFLUENCE: what winning looks like

Winning looks like a firm that knows its prompts, owns them and counts the result.

The Baltimore personal injury firm: +425% AI visibility growth, 56% on the top slip and fall prompt, 2 signed clients attributed to ChatGPT, 4 mo. time to attribution. From 5 AI leads and 0 signed.

WKND’s own proof:

  • Todd Miner Law, personal injury, Orlando (SEO). 8,166 ranking keywords; 7,500+ monthly organic visitors; 489 referring domains; #3 map pack; +288.9% keywords in six months, February 2025 to August 2025. Read the Todd Miner Law case.
  • A San Francisco law firm (ChatGPT Ads). “Google Search leads cost this San Francisco firm $248 each.” WKND’s page reports that “$11,982 in ChatGPT Ads spend produced 114 leads at $105.11 each,” 58% below Google Search, from June 15 to July 31, 2026. The result was judged “on intake quality and retained clients, not lead cost alone.” See ChatGPT Ads for law firms.
  • WKND’s AEO results: 156 snippets won, the average number of featured snippets acquired for clients within the first 6 months of AEO campaigns. 340% visibility increase, the average increase in SERP visibility through combined snippet, PAA, and AI Overview placements. 67% zero-click captures of target queries where clients’ content is displayed as the direct answer in search results.

I’ve built, scaled and sold three companies. The rule I run WKND on is the rule this Baltimore firm learned: count the signed client, not the click.

FAQ

Can a law firm get signed clients from ChatGPT?
Yes. A Baltimore personal injury firm tracked 2 signed clients to ChatGPT in LeadDocket, with 4 mo. time to attribution.

How long did the Baltimore firm take to get a signed client from AI?
4 mo. time to attribution, with +425% AI visibility growth across all four AI engines.

What did the firm do first?
It targeted the specific prompts Baltimore injury victims use, built bilingual authority with original Spanish-language signals, and put attribution on every client from first AI touchpoint through intake.

Do AI leads sign by default?
No. The firm’s first 5 AI leads came from wrong case types and jurisdictions and none signed. Prompt targeting and attribution turned AI visibility into signed clients.

How do I track signed clients from ChatGPT?
Use the AI Assistant channel in Google Analytics, call and form tracking by source, and one intake question: “How did you find us?”

Where do I start?
With an audit of your current visibility, the queries worth owning, and a scoped plan with timelines. WKND runs it free.

PERSUADE: get your own Baltimore play

Your city, your case types, your language market. WKND starts with “an audit of your current visibility, the queries worth owning, and a scoped plan with timelines. If the numbers do not support the work, we will tell you that instead of selling you a retainer.”

The price model: “We scope after an audit and price a fixed monthly program with no setup fee, so you see the full plan before committing.”

Why now: Google added a dedicated AI Assistant channel to Analytics on May 13, 2026, and the Search Console generative AI report reached every site on August 31, 2026. The tracking exists. The firms that name their prompts first take them first.

Get your free AI SEO audit for law firms, or call (916) 398-3999. Explore everything WKND runs on the services hub. For the signed-case math, read Is GEO Worth It for Law Firms?

WKND builds this for law firms through AI SEO for law firms and AEO services for law firms.

Related reading

Sources