Generative Engine Optimization (GEO): How Law Firms Get Cited in AI Answers

Generative engine optimization (GEO) is the practice of shaping your content, your website, and your presence across the web so that AI engines such as ChatGPT, Perplexity, Google AI Overviews, and AI Mode cite or recommend you in their answers. SEO earns a place in a list of links; GEO earns a place inside the answer itself.

For a law firm, the stakes are concrete. A prospective client who asks ChatGPT "who handles wrongful termination cases in Chicago?" gets a written answer that names two or three firms, often without clicking anything. If your firm is not one of them, it was never considered. Google is moving the same way: in our October 2026 analysis of 91 US search results pages on law firm marketing topics, 88 displayed an AI Overview.

This guide defines GEO, summarizes the research behind it, compares it with SEO and answer engine optimization (AEO), and lays out the six-step method we use with law firms.

What is generative engine optimization?

A generative engine is a search tool that writes its answer instead of returning ten blue links. It runs one or several searches, reads the pages it retrieves, and synthesizes a response with a large language model. Sources appear as numbered citations, as link cards, or sometimes only as paraphrased facts with no visible credit.

A definition you can use

Generative engine optimization is the set of techniques that make your content more likely to be retrieved, trusted, and cited by AI engines when they build an answer. The goal is not a ranking position but a share of the answer: being quoted, paraphrased, or recommended by name in the response the user actually reads.

Where the term comes from

The term was coined in a research paper, "GEO: Generative Engine Optimization", by Pranjal Aggarwal and colleagues from Princeton University, Georgia Tech, the Allen Institute for AI, and IIT Delhi. It was first posted on arXiv in November 2023 and presented at KDD 2024, a major data science conference. The paper treats AI engines as a black box and asks a simple question: what changes to a web page make an engine use it more?

What GEO covers in practice

Three conditions decide whether a model uses your page. Retrieval: the engine must be able to crawl your page and find it when it searches. Trust: the page must come from a source the engine can identify and verify, such as a named attorney at a real firm. Extractability: the page must state facts clearly enough that a model can lift them into an answer without guessing.

That is why we treat GEO as an extension of law firm SEO, not a replacement. Retrieval still runs largely on search indexes, so the SEO groundwork stays. GEO adds the work on trust and extractability that classic rankings never fully rewarded.

What the GEO research actually tested

Most GEO advice you will read traces back to this paper, so it is worth knowing what it found, and what it did not.

The setup: 10,000 queries and nine tactics

The researchers built GEO-bench, a benchmark of 10,000 queries drawn from nine sources and spread across 25 domains. They rewrote source pages using nine methods: authoritative tone, statistics addition, keyword stuffing, citing sources, quotation addition, easy-to-understand wording, fluency optimization, unique words, and technical terms. Then they measured how much of each answer drew on the rewritten page, weighted by position, plus a subjective impression score.

The three tactics that won: citations, quotations, statistics

Three methods stood out. Citing credible sources, adding quotations from relevant sources, and replacing vague statements with statistics produced a relative improvement of 30 to 40% on the position-adjusted word count metric and 15 to 30% on subjective impression. The abstract sums it up: GEO "can boost visibility by up to 40% in generative engine responses."

One finding matters directly to lawyers. The paper reports that the "Law & Government" domain benefits significantly from statistics addition. Data-backed legal content, such as deadlines, filing thresholds, or court statistics with their source, is the kind of content the models favored in these tests. Edits for fluency and readability also helped, by 15 to 30%.

What failed, and who gains the most

Keyword stuffing, the oldest SEO trick, offered little to no improvement. In a live test on Perplexity, it performed 10% worse than the unmodified page, while quotation addition improved visibility by 22% and statistics addition by up to 37%.

The most encouraging result concerns smaller players. Pages ranked fifth in the classic search results gained 115.1% in visibility with the cite sources method, while the same method cut the visibility of first-ranked pages by 30.3% on average. A solo practitioner or small law firm with well-sourced pages can win a place in the answer against bigger names.

Keep the limits in mind: the tests ran in 2023 and 2024, on a research engine and on Perplexity, and the engines have changed since. Treat the numbers as a direction, not a guarantee.

GEO vs SEO vs AEO: is GEO better than SEO?

GEO is not better than SEO, because it depends on it. The three disciplines target different surfaces of the same search journey, and a law firm needs all three working together.

Three disciplines, three goals

Answer engine optimization came first, with featured snippets and voice assistants: write one clear answer to one question. GEO keeps that habit and adds what generative engines need before they trust you: sources, data, and an author they can identify across the web.

Where the strategies diverge

In SEO, you optimize to be found. In GEO, you optimize to be understood and reused. That means fewer generic pages and more pages that answer one precise question, with a named attorney, a visible update date, and facts a model can quote word for word.

Rankings and citations do not always match. A competitor with an average website can be cited regularly if its pages are precise and attributable, and a firm that ranks well can be left out of the AI answer. That gap is the opportunity for firms that move early.

Is SEO dead now with AI?

No. Google's documentation says there are no additional requirements to appear in AI Overviews or AI Mode: a page must be indexed and eligible to show with a snippet, the same rules as classic search. If your pages are not indexed, most engines will never retrieve them. Google even says that clicks from results pages with AI Overviews are "higher quality," meaning visitors spend more time on the site.

How do AI engines choose their sources?

Every engine follows the same broad pattern: rewrite the question into searches, retrieve candidate pages, then let a language model pick the passages that best support an answer. The details differ by engine, and they explain most GEO tactics.

ChatGPT search

According to OpenAI's help center, ChatGPT search "sometimes partners with other search providers," Microsoft among them, and rewrites your question into one or more targeted queries, adding an approximate location for local searches. Its own crawler, OAI-SearchBot, collects pages for search answers, and OpenAI states that sites opted out of it will not be shown in ChatGPT search answers. Our guide on how to get cited by ChatGPT covers the tactics specific to it.

Perplexity

Perplexity searches the web as it answers and shows numbered citations that link to each source. Its crawler, PerplexityBot, is "designed to surface and link websites in search results on Perplexity," and the company says it is not used to train AI foundation models. Because every source is visible, Perplexity is the easiest engine to audit. More in our guide to Perplexity for lawyers.

Google AI Overviews and AI Mode

Google's AI features draw on its regular search index and use a "query fan-out" technique: they issue multiple related searches across subtopics and data sources to build one answer. That is why a page that answers a narrow sub-question well can be cited for a broad query. Google adds that no AI text file or special schema markup is required. See our guides to AI Overviews for law firms and Google AI Mode.

Across engines, a source gets cited when the engine can retrieve it, confirm who wrote it, and find a passage that answers the exact question on its own.

The GEO method, part 1: make your firm readable by AI engines

Our method has six steps in two blocks. The first block is technical: it makes sure AI engines can reach, read, and identify your firm. Without it, the content work in part 2 goes unseen.

Step 1. Set your robots directives for AI crawlers

Many law firm websites block AI bots without knowing it, through a security plugin, a CDN setting, or a copied robots.txt file. Each AI company runs separate bots for search and for model training, and you can treat them differently. The table below summarizes what each one does, according to each company's own documentation.

Blocking a training bot does not remove you from the matching search product: OpenAI's documentation states that each setting is independent, so you can allow OAI-SearchBot while disallowing GPTBot. Blocking a search bot, on the other hand, can take you out of that engine's answers. Allowing training bots may help models learn your firm's name over time; blocking them is a legitimate choice if you object to your content being reused.

Check your CDN and firewall too, since Google's guidance names hosting infrastructure as a place where crawling gets blocked. The Perplexity and Anthropic help pages give the exact user agents, also listed in our glossary entry on AI crawlers.

Step 2. Describe your firm and attorneys with structured data

Structured data will not get you into an AI answer by itself; Google says no special schema is needed. Its value is clarity: it tells engines that the Jane Doe on your bio page is the same attorney as the Jane Doe in the state bar directory. Use LegalService for the firm (schema.org now marks the older Attorney type as deprecated), Person for each lawyer, and Article for guides, with author and update dates.

The key property is sameAs. On each Person, list the URLs of that attorney's state bar profile, Avvo, Justia, LinkedIn, and any other directory listing. Keep the markup consistent with what the page visibly says, which Google lists among its best practices for AI features.

Step 3. Add an llms.txt file, without expecting miracles

An llms.txt file is a short Markdown file at the root of your site that gives AI agents a curated map of your most useful pages. Jeremy Howard published the proposal in September 2024. It is cheap to create and harmless. But Google states plainly that you don't need "AI text files" to appear in its AI features, so treat llms.txt as a convenience for agents that read it, not as a way to get cited.

The GEO method, part 2: give AI engines a reason to cite you

Once engines can read you, these three steps decide whether they choose you.

Step 4. Write answer-first pages that stand on their own

Models extract passages, not whole pages. Each section should answer one question in its first sentence, then explain. If a prospective client asks how long they have to sue after a car accident in Texas, your page should open with: "In Texas, you generally have two years from the date of the injury to file a personal injury lawsuit (Texas Civil Practice and Remedies Code § 16.003)." The details and exceptions come after.

Apply this to every practice area page and guide: one page per question, a direct answer under each H2, short paragraphs, and tables for comparisons. Add a visible "last updated" date. Bing's guidance on AI answers stresses that regular updates help AI systems reference the most current version of your content, and legal answers age fast.

Step 5. Back every claim with citations, quotations, and statistics

This is the step the GEO paper measured directly. Turn vague statements into sourced facts: cite the statute, the court rule, or the agency page; quote the exact text of the law or a court's holding; replace "many cases settle" with a number and its source. These edits take little time, and they match the three tactics that raised visibility by 30 to 40% in the research.

Legal ethics still apply to anything an engine might quote. Under ABA Model Rule 7.1, no communication about your services may be false or misleading, and Rule 7.2(c) limits who can claim to be a certified specialist. Present past results with context, and check your state bar's rules, as covered in our guide to lawyer advertising rules.

Step 6. Build your entity across the web

Engines trust what they can confirm in several independent places. For a US lawyer, the core profiles are the state bar's public attorney directory, Avvo, Justia, Martindale-Hubbell, FindLaw, Super Lawyers if you are selected, LinkedIn, and your Google Business Profile. Directories such as Avvo often list attorneys before they ever sign up, so you may already have a profile waiting to be claimed.

Then make the details match everywhere: firm name, attorney names, address and phone number (NAP), bar admissions, and practice areas, the groundwork of local SEO for law firms. Each attorney needs a real author page on your site with bar admissions, education, and links to those profiles, and every article should carry a byline that points to it. These are the E-E-A-T signals Google describes, and they show AI engines who stands behind the content.

Earned mentions complete the picture: a quote in a local newspaper, an article in a bar journal, a podcast interview, a CLE presentation. Run your target questions in Perplexity, note which third-party sites it cites for your practice area, and aim for a presence on at least one of them each quarter.

How to measure your AI visibility

No AI engine offers a full equivalent of Google Search Console yet, and answers vary by user, location, and day. The goal is not an exact score but a trend, and a list of blind spots to fix.

Build a prompt panel and test it monthly

List 10 to 20 questions your prospective clients actually ask, mixing informational questions ("how is child support calculated in Ohio?") with hiring questions ("best DUI lawyer in Phoenix"). Run them every month in ChatGPT, Perplexity, Google AI Mode, and Gemini, from a logged-out browser in your city. Record whether your firm is named, which page is cited, and which competitors appear instead.

Use the free data the platforms already give you

Google includes AI Overviews and AI Mode traffic in the overall Web search data of the Search Console Performance report. Bing goes further. Since February 2026, Bing Webmaster Tools has offered an AI Performance report, in public preview, showing how often your pages are cited in Microsoft Copilot and Bing's AI summaries, which URLs are cited, and the grounding queries behind them. Microsoft calls it an early step toward GEO tooling.

In GA4, set up an exploration or a custom channel for referrals from chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com. Part of this traffic arrives without a referrer and is counted as direct, so treat the figure as a floor. What matters for a law firm is whether these visits turn into calls and consultation requests.

What is the best tool for generative engine optimization?

No single tool sees inside every engine. For tracking at scale, Ahrefs Brand Radar monitors brand mentions and cited pages across AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini, and Copilot, and the Semrush AI Visibility Toolkit tracks how brands appear in AI answers, including ChatGPT and Google AI Mode. For most small firms, a monthly prompt panel plus Bing's free report gives enough signal to act.

A minimal GEO tracking protocol in four steps

1. Write 10 to 20 prospective-client questions, with your city where relevant.
2. Test them monthly in ChatGPT, Perplexity, Google AI Mode, and Gemini, and log which firms and pages are cited.
3. Check the Bing Webmaster Tools AI Performance report and your GA4 AI referrals each month.
4. List the three third-party sites cited most often in your practice area and earn a presence on one of them this quarter.

GEO for law firms: what it looks like in practice

Small and midsize firms are well placed to benefit. National directories dominate classic rankings through sheer volume, but a precise, well-documented firm rooted in one city can earn its place in AI answers on high-value questions.

A worked example

A business owner asks Perplexity: "Which lawyer in Austin can help me sell my small business?" The answer cites two or three sources. A firm with a dedicated page on selling a business, a byline from a named attorney admitted in Texas, matching Avvo and state bar profiles, and a mention in a local business journal has a real chance of being cited. A firm with none of those signals will not be, however well it ranks.

Matching AI queries with content

The payoff is the ratio of effort to visibility. Five strong, well-sourced pages on your core practice areas, backed by a consistent entity across the web, keep working long after a paid campaign stops. That is what we see at Hendy with the firms that built these signals early.

What generative engine optimization services should include

If you hire an agency for GEO, judge the offer on method and measurement, not on buzzwords. A serious engagement covers both the technical side and the editorial side.

The deliverables to expect

• A crawl audit of robots.txt, CDN, and firewall rules for AI bots.
• Structured data for the firm and each attorney, with sameAs links.
• A rewrite plan for priority pages: answer-first sections, citations, statistics.
• An entity cleanup: consistent NAP, claimed profiles, author pages.
• A prompt panel and monthly reporting on citations, mentions, and AI referral leads.

Red flags

Be wary of anyone who guarantees citations in ChatGPT or a spot in AI Overviews: no one controls these answers, just as no one can guarantee the top position on Google. Be equally wary of mass-produced AI content and of offers that sell GEO apart from SEO. The two share the same foundation, which is why our law firm SEO services handle both together.

Key takeaways

AI engines do not simply cite the pages that rank best on Google. They cite the sources they can retrieve, identify, and quote with confidence. For a law firm, generative engine optimization comes down to six steps: open access to AI search crawlers, add structured data, publish llms.txt as a small extra, write answer-first pages, source every claim, and keep your entity consistent across the web. Then measure every month.

Firms that start now take a lead that is hard to close, because the pages, profiles, and mentions behind a citation take months to build. To learn where your firm stands in AI answers and what to fix first, talk to our team. For the wider picture, read our guides to AI for lawyers and AI search optimization for law firms.