A quiet shift is underway in how prospective clients find legal help, and most law firms have not adjusted to it yet.
For two decades, "how do clients find a lawyer?" had a simple answer: they typed something into Google and clicked one of the top results. That is still true for the majority of legal searches today — but no longer the only path. A growing share of prospective clients now ask ChatGPT, Perplexity, Claude, Google's AI Overviews at the top of the search results, or the AI assistant baked into their phone. Instead of a list of ten blue links, they get a paragraph. That paragraph either mentions your firm or it does not.
This guide explains what is changing, why it matters specifically for law firms, and what your firm can practically do about it without falling for hype.
## What "AI search" actually refers to
The phrase covers four overlapping experiences that look different but share the same underlying mechanic.
1. Conversational AI assistants. ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot. The user asks a question in plain English; the assistant returns a written answer, sometimes with citations to source websites. People increasingly ask these tools for recommendations: "Who is a good estate-planning attorney in Austin?", "What should I do if I get a DUI in Maryland?", "Find me a family law firm that handles complex custody cases in Cook County."
2. AI Overviews on Google. Google now displays an AI-generated summary at the top of many search results pages, particularly for informational and comparative queries. The summary pulls from a small number of sources — usually three to seven — and links back to them. Sites that get pulled into the Overview win significant visibility; sites that do not get pulled in lose it.
3. AI-powered features inside the existing search engines. Bing's Copilot sidebar. DuckDuckGo's AI assist. The voice assistant on the user's phone reading back a single answer instead of returning a list. All of these compress ten results into one result, and that one result is dramatically more valuable than any of the ten used to be.
4. AI agents that complete tasks on the user's behalf. Still early, but growing fast. "Schedule a consultation with a personal injury attorney near me who handles motorcycle accidents and offers free initial consultations." The agent reads firm websites, evaluates them against the criteria, and either books or shortlists. Firms whose websites are structured for machine reading get represented; firms whose websites are structured only for humans get skipped.
Each of these surfaces is sometimes called generative search or answer engine search. The discipline of optimizing for them goes by names like AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization). Different label, same underlying goal: be the source the AI references.
## Why this matters specifically for law firms
Three reasons AI search matters earlier and more sharply for legal services than for many other industries.
The questions prospective clients ask are exactly the kind AI is good at answering. "What should I do after a slip-and-fall accident?" "How is property divided in a Texas divorce?" "Do I need a lawyer for an uncontested guardianship?" These are informational questions with relatively stable answers. They are precisely the queries where AI tools confidently produce summaries — and precisely the queries that used to drive prospective clients to your blog content. That traffic is now partially routed through the AI.
The recommendation question is now being asked of AI directly. When someone asks ChatGPT "Who is a good probate lawyer in Phoenix?", the assistant pulls from publicly available signals — your website content, your reviews, your media mentions, your professional directory listings — to produce its answer. If your firm is well-represented in those sources, you can end up named in the answer. If not, your competitors will be.
Trust signals are weighted even more heavily than in regular search. AI tools err on the side of caution for high-stakes topics like legal advice. They lean toward citing sources that look authoritative — bar association directories, established legal publications, firm websites with clear attorney credentials, content with properly formatted citations and dates. Generic, weakly attributed content gets passed over. Strong, expertise-signaled content gets surfaced.
In other words: the playbook that wins traditional SEO for law firms — depth, accuracy, jurisdictional precision, attorney expertise on the byline, structured data, and reputable third-party validation — is the same playbook that wins AI search. The good news is that nothing radical has to change. The bad news is that firms that have been cutting corners are about to find that the corners they cut now matter more, not less.
## How AI tools decide what to cite
You do not need to understand how the underlying models work. You do need to understand what they look at when answering a query about your area of practice or your geographic market.
Source authority. Is this site one that the model has seen treated as authoritative across many other sources? Bar association sites, established legal publications, large news outlets, and recognized legal directories all carry weight. So do firm sites that have themselves been cited by those authoritative sources over time.
On-page expertise signals. Is the article or page bylined by a named, credentialed attorney? Are credentials, jurisdictions, and dates clearly visible? Is there a date stamp showing when the content was reviewed? Is there a fact-checked-by or reviewed-by attribution?
Structured data (schema markup). Is the page wrapped in machine-readable structured data identifying the firm, the attorneys, the practice areas, the office locations, the hours, the licenses? AI tools use schema to confirm what a page is about without having to guess. Pages with rich, accurate schema get cited more often.
Topical depth. Does the site cover the topic comprehensively, or is there one shallow blog post on it? AI tools favor sites that have demonstrated subject-matter coverage — five to twenty interconnected pieces on a topic outweigh one stand-alone post.
Reputation signals. Reviews, ratings, mentions in the news, citations in legal directories. AI tools fold these into their judgment about whether a firm is a credible recommendation in a specific market and practice area.
Content readability for a model. Pages that load quickly, render their main content in HTML rather than only in JavaScript, use clear headings, write in plain language with
direct answers, and avoid burying the lede in marketing fluff are easier for AI tools to extract from. Pages that hide the actual answer behind layers of stock copy get passed over.
## Where most law firms are vulnerable right now
Five gaps we routinely see when auditing law firm websites against AI search readiness.
1. No author bylines or weak ones. Articles attributed only to "Admin" or "The Firm" or with no byline at all give AI tools nothing to verify. Articles bylined to a specific named attorney with credentials and a linked bio pass the expertise check.
2. Stock content with no jurisdictional specificity. Generic articles that could apply to any state get filtered out in favor of articles clearly written for the searcher's jurisdiction. "Personal injury statutes of limitations vary by state" is a hedge; "In California, the statute of limitations for personal injury is generally two years from the date of injury" is an answer.
3. No or thin schema markup. Many firm websites have either no structured data or minimal structured data — usually because the agency that built the site never set it up. AI tools cannot confirm what they cannot read.
4. No clear expertise signals on practice-area pages. A page titled "Estate Planning" with three paragraphs of marketing copy and no attorney attribution loses to a page titled "Estate Planning Attorney in Houston, Texas" with named attorneys, listed credentials, jurisdictional context, and a substantive FAQ section.
5. Slow, JavaScript-heavy pages that are hard to read. Some firm websites render almost everything through JavaScript that some AI crawlers handle imperfectly. The content is technically there; the model cannot easily see it.
## What you can do, in plain terms
Some of this is technical, but the practical actions are easy to describe.
### Strengthen the bylines
Every meaningful article on your firm's website should be bylined to a specific, named attorney. The byline should link to a full attorney bio that lists their bar admissions, years of practice, education, and the practice areas they handle. This is the single highest-impact change most firms can make for AEO and is also good for traditional SEO and for state-bar compliance.
### Add a "reviewed by" or "last updated" stamp
Visible dates and review attributions help AI tools assess freshness and accountability. "Reviewed by [Attorney Name] on [Date]" placed prominently near the top or bottom of an article is a strong signal.
### Be jurisdiction-specific in every legal article
If an article references a statute, a procedure, a deadline, or a damages cap, it should name the state (and county, where relevant). Stock content rewritten for your state — with your state's actual numbers, deadlines, and procedures — is exponentially more useful and more likely to be cited.
### Audit your structured data
Most law firm websites need a baseline of structured data: LegalService, Attorney, LocalBusiness for each office, FAQPage on FAQ-style content, Article on blog posts, and the appropriate review markup where relevant. (Our separate Schema markup for law firm websites guide walks through this in plain English.)
### Build out topical depth
Pick two or three of your highest-priority practice areas and commit to building a real cluster of content around each one — a strong main practice-area page, multiple supporting articles answering specific questions clients ask, an FAQ section, and well-structured internal linking between them. AI tools strongly favor sites that can be confidently identified as authoritative on a specific topic, not sites that have one paragraph on each of fifteen topics.
### Make sure third-party signals are in good shape
Confirm your firm's name, address, and phone number are consistent across the major legal directories (Avvo, Justia, FindLaw, Lawyers.com, your state bar). Encourage genuine, ethical reviews on Google. Pursue the legitimate media mentions and bar-association recognition that demonstrate your firm exists in the real world, not just on its own website.
### Test what AI says about your firm today
Open ChatGPT, Perplexity, and Google's AI Overview, and run the queries a prospective client would run. "Who is a top family lawyer in [your city]?" "How do I find a [practice area] attorney near [your city]?" See whether your firm is mentioned, mentioned accurately, or omitted entirely. Run the same queries every quarter. The picture changes faster than annual marketing reviews can keep up with.
## What you should not do
A few things to avoid, because they appear in agency pitches but they do not work and may actively hurt you.
- Do not stuff AI keywords into your pages. "Best AI-recommended attorney near me" is not a real query and will not help. AI tools are trained to ignore that kind of manipulation.
- Do not generate volumes of AI-written content with no expert review. Bulk AI-written legal content that has not been reviewed by a licensed attorney in the relevant jurisdiction is both a state-bar compliance risk and a quality risk that AI search engines themselves are getting better at detecting and demoting.
- Do not pay vendors who promise to "get you into ChatGPT's recommendations." No one has direct paid placement in those answers today. Anyone offering it is either misunderstood or dishonest.
## How AI search and traditional SEO fit together
The most important thing to understand: AI search has not replaced traditional Google search. Google still drives the majority of legal traffic, and the local map pack still drives the majority of local-intent legal searches. AI search is additive — a growing share of attention that sits on top of the traditional channel. Both matter, and both are won with substantially the same playbook.
Firms that wait until AI search is the majority channel will be too late. Firms that overinvest in AI search at the expense of fundamentals (local presence, reviews, technical SEO, content depth) will be wasting money. The right answer for almost every firm right now is: keep doing the traditional SEO work seriously, and use the same work to also satisfy what AI tools look for. There is more overlap than there is divergence.
## How
LexGrowSEO approaches AI visibility
The LexGrow SEO program already builds for this future as part of the standard work — bylined attorney content, jurisdictional precision, schema markup on every key page, topical clusters built around your practice areas, and ongoing measurement of how your firm appears across both traditional and AI search surfaces. The dashboard tracks where you appear in traditional rankings and surfaces visibility across AI Overviews and conversational answer engines for the queries that matter most in your market.
You do not have to become an AI expert to be visible in AI search. You have to commit to the kind of substantive, expertise-signaled content marketing that wins regardless of which surface a prospective client uses.
## The next decade of legal marketing
For 25 years, the playbook for being found online has been "rank on the first page of Google." For the next 25 years, the playbook will be "be the answer." Sometimes that answer will be served as a list of ten blue links, sometimes as a paragraph in an AI Overview, sometimes as a single recommendation from a chatbot, sometimes by an autonomous agent acting on behalf of a prospective client.
Your firm's job is not to chase every new surface. It is to invest in the underlying assets — depth, expertise, reputation, accuracy, accessibility — that make you the credible answer no matter how the question gets asked.
Firms that do this win the next decade. Firms that wait for the surface to settle will find that, by the time it does, the leaders have already been chosen.