A few years ago, someone searching for a local attorney opened Google, typed "family lawyer near me," and started scrolling. They would skim the map pack, click two or three firms, open reviews in a second tab, then pick up the phone. The search engine gave them a field of options and let them sort it out themselves.
That same person today opens ChatGPT or Gemini, types a sentence about their situation, and gets back a short answer with two or three firms already named. No map. No scrolling.
The sorting happened before they ever saw it. For the professional on the other end of that call, the story is simple: either your name was in the answer, or the prospect never knew you existed.
The Shortlist Is Shorter Than the Map Pack Ever Was
The old local pack showed three businesses. The AI shortlist is often two. One researcher who pulled apart ChatGPT's local search behavior found a configuration setting that caps results at two businesses per query, a tighter squeeze than anything Google's map ever showed.
And getting into those two slots is harder than getting onto the map. SOCi's 2026 Local Visibility Index, reported by the National Law Review, looked at 350,000 business locations and found ChatGPT recommends only 1.2% of them, versus roughly 36% for Google's local pack. The AI assistants are something like 30 times pickier. For a deeper look at how AI answers are rewriting the way prospects find lawyers, it's worth reading alongside this piece.
Ranking Well on Google Doesn't Carry You Over
The uncomfortable finding is that firms winning in traditional local search are not the same firms the AI picks. The same SOCi study found that fewer than half of the businesses doing well in the map pack show up in AI recommendations at all.
The two systems weigh different signals. Google's local algorithm leans on proximity, Google Business Profile completeness, review volume, and on-page relevance. The AI assistants read differently. They pull from a wider pool, lean on text they can summarize, and seem to care more about what independent sources say about a business than what the business says about itself. A firm that spent five years polishing a Business Profile and chasing five-star reviews may still be invisible to the shortlist.
Who Actually Lands in the Two Slots
The firms that keep showing up in AI answers tend to share a few traits. None are secrets, and none are quick fixes.
- A substantive website. The assistants pull heavily from a firm's own pages when they summarize who does what. Pages that answer specific questions in plain language give the model something to lift.
- Mentions on sources the model trusts. Being named in reported articles, bar association listings, and legitimate directories matters more than it used to.
- A real presence in community discussions. Reddit threads, local forums, and Q&A sites feed the models. A firm that has been genuinely useful in those places tends to surface.
- Reviews that read as authentic. Volume still matters, but so does the texture of the reviews and whether the firm bothers to respond.
The Prospect Shows Up Already Half-Decided
The second-order effect of the shortlist is what it does to the person who calls. They did not compare ten firms. They compared two. And they arrived with the assistant's framing of what their situation is and what questions they should be asking.
Pew Research's 2026 report on AI use found that 49% of U.S. adults now use AI chatbots, and information search is the top use case. For many practice areas, the intake call is now the second step in the prospect's research, not the first. If the assistant's summary is wrong, the firm has to correct it quickly without making the caller feel foolish. If it's right, the firm has to confirm it and move the conversation forward.
What to Change This Quarter
- Rewrite the service pages so they answer real questions. Replace generic practice-area copy with pages that address the specific situations prospects describe to the assistant.
- Earn mentions off your own site. Pitch reporters, contribute to bar publications, show up in legitimate directories, and participate honestly in community forums.
- Audit what the assistants currently say about you. Ask ChatGPT, Gemini, and Perplexity the questions your prospects ask, and see which firms get named.
- Train intake for the coached prospect. Give the team a short script for confirming what is accurate, correcting what is not, and moving to a consultation without condescension.
The shortlist is not going to get longer. As more people get their first answer from an assistant instead of a search page, the two names in that answer will carry more weight, not less. Being one of them is the work.
