Firms that AI recommends convert at 4.4x the rate of traditional SEO traffic (Semrush, June 2025). The client isn’t shopping; they’re being told who to hire. I build the entity data that forces ChatGPT, Gemini, and Perplexity to cite you by name.
90% of agencies are still fighting for blue links that now sit below the ads, below the Map Pack, and below the AI Overview. Traditional SEO competes for a slot on a list. Generative Engine Optimization competes to be the one firm the AI mentions in its answer.
Think Impression Share, but for AI. If “Miami medical malpractice” generates 100 AI queries per month and your firm appears in 40 of those answers, your SoMV is 40%. Most firms I audit are sitting at 0% and don’t know it yet.
Three numbers explain why firms relying on blue-link SEO are watching their pipeline dry up, even when they rank #1. All three are real and citable.
Pew Research tracked 900 US adults’ real browsing in March 2025. When an AI Overview appears, the click-through rate falls from 15% to 8%: roughly half. The #1 result on Google is now the AI, not your website.
ChatGPT weekly active users grew 8x in 18 months (100M in October 2023 to 800M in April 2025). A growing share of your prospective clients are bypassing Google search entirely. The number keeps climbing each quarter.
Users read the AI answer and leave. If you’re not cited in that answer, you don’t exist in their consideration. Zero-click searches went from 56% to ~60% since AI Overviews launched (Similarweb, 2025).
AI can only recommend a firm it can recognize as a distinct, verified entity. If your firm isn’t built that way, AI doesn’t even try. Three technical layers decide whether you get cited or skipped.
AI needs to know your firm exists as one verified entity, not a scattered set of mentions. I build the same firm profile across Wikidata, Crunchbase, bar directories, legal directories, and Google Business Profile, then connect them with SameAs schema. AI reads that as proof you are who you say you are.
AI weights the first sentence of your content most heavily. Lead with the answer and your firm’s name reads as the authority. Pages that open with three paragraphs of disclaimers get skipped entirely. The page above the fold is the page AI sees.
Every legal claim on the page needs to point to a .gov or .edu source. That creates a three-part structure AI stores and retrieves later: who your firm is, what you practice, and what statute governs it. The exact terms clients use in reviews feed back into schema so AI links your firm to real client language.
AI doesn’t read your website like a human. It breaks every statement into a three-part structure and stores it in a vector database. Get the structure right and AI retrieves your firm when anyone asks.
When a client asks ChatGPT “who handles Florida comparative fault medical malpractice?“, the AI pulls up this exact three-part chain. The firm is the subject. The practice is what they do. The statute is the law that governs it.
No chain, no citation. Generic content reads well to a human and tells AI nothing. Only content with named entities, defined practices, and source-linked statutes gets stored in the way AI can retrieve later.
For a med mal firm, the buildout produces hundreds of these chains. Each one links the firm name to a specific medical condition, a specific Florida statute, and a verifiable source on a .gov or .edu domain.
AI tools like Perplexity and Gemini don’t only read what they were trained on. They browse the live web while answering a question, then pick which sources to cite. Your firm needs to show up across the exact directories and publications AI checks in real time.
As the user types, AI scans bar directories, court records, legal publications, and news mentions. It doesn’t only check your website. It checks the third-party sources where your firm should already appear, then decides whether to recommend you.
When your firm name shows up next to the same practice-area words across many different sites, AI starts treating that as proof you’re the authority. The more sources, the stronger the link. This is how AI separates a real expert from a generic listing.
Standard SEO breaks down in South Florida because every firm chases the same keywords with the same content. I worked around it by building entity signals AI uses instead. The proof shows up in a language managing partners understand: named recommendations from AI that turn into signed cases.
Percy ranked well in Google, but when I tested AI searches, the answers came back generic: “contact multiple lawyers, get free consultations.” His name wasn’t in any of them. Share of Model Voice was 0% when I started.
I pulled the exact phrases real clients used in reviews. When a client writes “Percy handled my surgical error case,” AI reads that as a real-world expertise signal stronger than anything we could write ourselves. 400+ reviews and 10+ legal directories got linked through SameAs schema as one verified entity.
Percy’s name now appears next to “surgical error” and “Miami” across enough independent sources that AI treats him as the default specialist. Cited by ChatGPT, Gemini, and Perplexity by name within 4 months.
Jorge hit #1 in the Kendall Map Pack in 5 months. Most agencies would call that a win and stop there. I treated it as the training data for the AI layer.
I synced his Google Business Profile with the right Geo Schema markup. That tells AI tools like Perplexity that Jorge is the verified firm for the Kendall submarket specifically, not one of many Miami options. The Map Pack data feeds the AI’s local layer directly.
As AI search keeps growing, Jorge is already locked into the local data AI pulls from. When someone asks Gemini “who handles Kendall birth injury,” his firm is the answer.
Every deliverable is technical, verifiable, and tied to how AI evaluates authority. Here’s what lands in your account, in plain English.
Every audit starts with me asking ChatGPT, Gemini, Perplexity, Claude, and Bing Copilot if they recommend your firm for your practice area in your market. Here’s what a scorecard looks like for a typical Florida firm before I start.
The scorecard runs 5 platforms, 5 queries, real-time citation checks. You get a number, a breakdown by platform, and the specific gap that’s keeping competitors in the answer and you out of it.
Get my scorecardTraditional SEO fights for a position on a list of links. AI search optimization (also called Generative Engine Optimization, or GEO) fights for a citation inside the answer that ChatGPT, Gemini, or Perplexity generates.
The mechanism is different. Instead of keywords and backlinks, AI evaluates entity recognition, authority signals, and whether your content is structured for machine extraction. You still need a fast site and quality content, but the additional layers are what determine whether AI cites you or your competitor.
Ask it. Type your practice area and city into ChatGPT and Perplexity and see if your firm appears. Try “best personal injury lawyer in Tampa” or “medical malpractice attorney Sarasota” or “criminal defense lawyer Fort Lauderdale.”
If your firm isn’t in the response, your entity data isn’t strong enough for AI to recognize and recommend you. The AI Visibility Scorecard runs this test across all major platforms and shows you where you stand versus competitors in your market.
No, it builds on top of it. You still need a fast website, quality content, and local signals. Traditional SEO alone won’t get you cited in AI answers.
Entity optimization and answer engine formatting are the additional layers that determine whether AI recommends you or your competitor when someone asks for a lawyer. The two engagements are scoped separately and priced separately; many firms run both with the same operator (me) for tracking consistency.
Most firms see partial AI citations within 60 to 90 days as entity data propagates through Wikidata, bar directories, and Google’s Knowledge Graph. Full Share of Model Voice above 30% typically takes 4 to 6 months.
It’s slower than traditional SEO in the sense that you can’t trick it; you have to build the entity. Once it’s built, it’s harder for competitors to knock you out because they’d have to replicate the same multi-source verification pattern.
Semrush published a June 2025 study analyzing 500+ digital marketing and SEO topics. AI-recommended visitors converted at 4.4x the rate of traditional organic traffic because they’re not comparison shopping.
A client who asked ChatGPT “who’s the best med mal lawyer in Miami” already accepted the AI’s answer as a recommendation. By the time they hit your site, they’re ready to call. They’re not clicking through ten blue links and forming opinions; they’re being told who to hire.
Yes. Miami is one of the most competitive legal markets in the United States. If the protocol works here, it works in Austin, Dallas, Atlanta, or anywhere else.
The entity buildup process is market-agnostic; what changes is the specific directories and local publications I target. A Texas firm gets Texas Bar, Texas Lawyer magazine, and Texas court records instead of the Florida equivalents. Same engine, different inputs.
I’ll run your firm through ChatGPT, Perplexity, Gemini, Claude, and Bing Copilot to see if they know who you are. You get a scorecard showing which platforms cite you, which cite your competitors, and where the gaps are. If AI doesn’t know you exist, I tell you exactly how to fix that.
