Generative engine optimization (GEO) for law firms: how to get cited in AI search in 2026

Most partners we talk to underestimate how often this happens. When SE Ranking looked at legal search terms, 77.67% of them returned a Google Overview. Traffic from AI assistants is still small, but it's climbing fast. In the Previsible LLM Traffic Report, ChatGPT and similar tools accounted for 0.37% of sessions on legal sites in December 2024 and 0.86% by March 2025.

This guide walks through what to do about it, starting with how LLM picks its sources. We've also written about AI search optimization in other industries.

What is generative engine optimization for law firms?

Generative engine optimization for law firms is the process of making your practice easy for AI assistants to recognize, describe correctly and recommend. These tools write their answers from material they've found online. A firm with clear pages and plenty of outside mentions gives them something to work with. A firm with a five-page website and a stale Avvo profile usually gets left out.

A Google search gives a person ten links to choose from. Ask an AI tool for a good workers' comp lawyer in Columbus, though, and you get a paragraph with a couple of names in it. Plenty of people just call one of them.

The main tools to watch are ChatGPT, Perplexity, Gemini, Claude, Bing Copilot and Google's Overviews. Perplexity is the most transparent of the group and lists the pages it used under each answer. ChatGPT usually doesn't bother with links. The reader gets a name and has to search for the firm on their own. With AI Overviews, what you see usually comes from pages that already rank well in Google. A good plan accounts for these differences.

Why this matters in 2026

People usually come to a lawyer worried and a bit confused. An AI assistant lets them work out what's going on in private at 11 p.m., and by the time they're ready to call a firm, it may already have suggested one.

Searches like "how is fault decided in a rear-end collision" used to send people to attorneys' blogs. Now Google often answers at the top of the page, so your article can rank first and still get fewer clicks than it did two years ago.

There's also a timing issue. In most cities there's no clear AI favorite for each practice area yet. Once ChatGPT settles on, say, one personal injury firm in Denver, it keeps reusing the same sources, and pushing that firm aside takes far more work than getting there first.

How LLM decides who to cite

None of the AI companies publish what they look for, but when we compare who gets mentioned with who doesn't, a few things keep coming up.

The first is identity. If you're "Smith & Lee LLP" on your website, "Smith Lee Legal" on Yelp and "Smith and Lee Attorneys" on Avvo, a model may not connect them. The same goes for old addresses and phone numbers.

Authorship matters a lot with legal topics, which Google treats as "Your Money or Your Life" content. A page about Florida's statute of limitations signed by an attorney, with a bio that lists bar admissions, is a much safer source than an unsigned blog post.

The content itself has to be quotable. If someone asks "how long do I have to file a personal injury claim in Florida?", the page that says "two years for most negligence claims since the 2023 reform" in its first line is far more useful to an AI than one that opens with three paragraphs about the firm's values.

Then there's what other people say about you. Reviews, a quote in a local news story, a listing with the state bar, an article in a legal publication. Perplexity makes this easy to see because it shows its sources, and our piece on Perplexity AEO goes into how it picks them.

Finally, there's the technical side. If your robots.txt blocks GPTBot, OAI-SearchBot, PerplexityBot or ClaudeBot, those tools may not be able to read your site at all. We still find this on attorney websites, usually left over from a blanket "block AI" setting someone turned on in 2023.

Core strategy, step by step

There are six parts to the work. The first two or three usually pay off soonest, so we've put them at the top.

Structured data and schema markup

Behind most well-built legal websites there's a block of code called schema markup. It spells out basic facts about your practice for search engines and AI tools. The type to use is LegalService, and each office gets its own entry. Attorneys get Person markup that links to their bar profile and LinkedIn. If you have Q&A pages, add FAQPage too. You may see the older Attorney type recommended in guides. Schema.org has deprecated it in favor of LegalService.

One catch with FAQ markup is that Google stopped showing FAQ dropdowns for most sites in 2023, and legal sites aren't among the exceptions. It's still worth adding. Any developer can set it up as JSON-LD in an afternoon, and Google's Rich Results Test will tell you if something's broken.

Content structuring for citation

Practice area pages tend to read like brochures, which is fine for a visitor but not much use to an AI tool looking for an answer. Your intake staff can help here. They know the ten questions people ask most in each practice area. Turn each one into a heading and get an attorney to answer it straight away, then explain the details below. Nobody types "Our Employment Practice" into ChatGPT. Your heading should be the question people really ask, something like "Can I sue my employer for wrongful termination in California?"

Authority and citation building

LLM have favorite sources, and finding out which ones takes about two minutes. Run a question about lawyers in your city through Perplexity and scroll to the sources. Avvo and Justia tend to show up, along with bar association pages and the odd Law360 article or local news story. A mention in one of them usually does more for you than a stack of links from random directories.

Getting into those places is mostly old-fashioned legal PR. Your attorneys probably do some of it already, whether that's talking to a reporter about a local case or writing for the bar journal. The difference now is knowing where to pitch. Go after the outlets AI tools already quote for your practice area before anything else.

Review and reputation management

For a question like "a good divorce lawyer near me," AI tools lean heavily on reviews. A firm with forty reviews, some of them recent, has an edge over one with six from 2021. The wording counts as well. A review that says "they handled my custody case in Austin" tells the AI more than "great service!"

Ask for reviews as part of closing a matter, and spread them across Google, Avvo and Lawyers.com. Check your state bar's rules on testimonials first. They differ quite a bit from state to state.

Practice-area content clusters

A single "Personal Injury" page doesn't give an AI much to go on. A dozen pages that each cover one piece of it, say car accidents, slip-and-fall claims or your state's filing deadline, show you actually know the area.

It makes sense to start where the money is. A family lawyer might start with custody. A criminal defense firm could start with DUI, since the questions are common and the answers vary by state.

Technical setup

On the technical side, the usual problem is a robots.txt rule blocking crawlers. An outdated sitemap or pages that only load through JavaScript can cause trouble too. Our GEO services cover this side of the work if you don't have a developer on hand.

GEO vs. SEO: what's actually different?

If your firm has ever hired an SEO agency, you've seen the monthly reports. They're mostly about keywords and backlinks, and the goal is a higher spot in Google. GEO is less about position. What counts is whether an AI tool trusts your firm enough to use it in an answer, which depends on your reputation online and how clearly your pages explain things.

There's a lot of overlap. AI Overviews often cite pages from the first page of results, and things like attorney bylines tend to help rankings too. In most cases, generative engine optimization (GEO) for law firms in 2026 builds on SEO you already have.

The measurement is different, though. An SEO report shows positions and clicks. Here, you're checking whether LLMs mention you at all, what they say and which competitor they name when they don't name you. Your own site is also only part of the picture. Sometimes a good Avvo profile and a mention in the local paper do more for you here than a whole new section of your website.

How to measure results

The simplest way of measuring generative engine optimization for law firms is a spreadsheet of prompts. Ours usually has 20 to 50 rows, things like "best estate planning attorney in Phoenix" or "what happens at a first DUI hearing in Ohio." Every month someone runs them through ChatGPT, Perplexity, Gemini and Google and logs who gets named.

Three or four months of that and the trend is usually obvious. It's also how you find mistakes, like ChatGPT still listing an office you moved out of. For traffic numbers, GA4 is enough. AI visits show up under source/medium as Chat GPT, Perplexity and similar. It's worth watching how those visitors convert. In our experience they often arrive knowing what they need.

Doing this by hand is fine for a month or two. After that, a tracking tool saves a lot of copy-pasting.

Pricing changes a lot in this space, so check the vendor's site before you sign up.

A 30/60/90-day plan

Here's roughly how the first three months go when we take on a new legal client.

Month one is mostly checking. Run the prompt list to see which competitors AI keeps naming, then clean up any directory listings with outdated details. Add LegalService and Person schema to the site, plus FAQPage markup on any existing Q&A pages. Also get AI referral tracking working in GA4 now. You'll want the before numbers later.

The second month is when the writing starts. Rewrite your two or three most important practice area pages so they answer common questions up front, with a named attorney as the author. Start a set of question-based pages around your main practice area. This is also the month for citation outreach. Begin with the directories and local outlets that tools already cite for your competitors.

From day 61 to 90, roll the same approach out to the rest of your practice areas. Set up a routine for asking clients for reviews. Put together a simple monthly report and look at what's changed. By this point, you'll usually start appearing for at least some of the less competitive prompts.

Once the 90 days are up, there's less to do. Most of our clients rerun their prompts every three months or so and add pages when a new question keeps coming up at intake.

Do it yourself or hire an agency?

Plenty of lawyers handle this in-house. If you have one or two practice areas and someone in the office who understands basic SEO can spend a few hours a week on it, the plan above will get you started.

Outside help makes more sense when there's more to manage. A firm with four offices has four sets of listings to keep in sync, and personal injury in a big city takes real PR work to break into. And in a lot of offices nobody's really in charge of marketing, so this keeps getting pushed to next week.

If you're talking to GEO agencies for law firms, ask them to run a prompt for one of their clients while you watch. Ask whether they've worked with lawyers before and whether they know the attorney advertising rules in your state. A good agency should be able to tell you exactly what it tracks and show you a sample monthly report. Budgets vary a lot. A small single-office firm might spend around $1,500 a month, while a multi-office firm in a competitive market can easily spend $5,000 or more. You can read about how we work on our B2B GEO agency page.

Recommended articles

• GEO SEO for Local Businesses