Run the test before you read another paragraph.
Open ChatGPT. Type: “who are the best [your practice area] attorneys in [your city].” Read the answer. Then ask Gemini the same question. Then Perplexity. Note which firms get named. Note whether yours is one of them. If you’re like most managing partners we’ve put through this exercise, you’ll close the laptop with a different opinion of your market position than you started with.
A firm we recently spoke with had been in practice for 22 years. Six hundred plus five-star reviews. Top three Google rankings for every keyword that mattered. ChatGPT named four other firms when asked who the best car accident attorneys in his city were. None of them outranked him on Google. Two had fewer reviews. The conversation that followed started with: “I didn’t know this was something I needed to worry about.”
It is. BrightLocal’s 2026 Local Consumer Review Survey reports that 45 percent of consumers now use AI tools to find local services. The same study reported 6 percent the prior year. That’s a 7x change in twelve months in how potential clients start looking for an attorney. Meanwhile, Google’s AI features are quietly reducing the surface area of the local pack itself. Sterling Sky’s analysis of AI-powered local packs shows them displaying about a third as many unique businesses as traditional 3-packs across hundreds of markets, often with the call button replaced by an image carousel. Local pack ads grew from under 1 percent of mobile queries in November 2025 to roughly 22 percent by January 2026 according to Places Scout data.
The shift isn’t theoretical. It’s measurable, and it’s accelerating quarter over quarter.
Ranking is not the same as being cited
Traditional SEO is a positioning problem: you compete for a slot in a ranked list and your job is to be higher than the firm beside you for a given keyword. AI search is different in kind, not in degree. There’s no ranked list, and the only question is whether the AI has enough confidence in your firm to name you at all. When someone asks Perplexity who handles wrongful death cases in Houston, the answer isn’t ten firms in a ranked order. It’s a generated response that names one firm, sometimes two or three.
That confidence is the entire mechanism. AI systems default to entities they can corroborate from multiple independent sources. When the signals around a firm are sparse, conflicting, or only present on the firm’s own website, the system reaches for a more recognizable entity instead. Your competitor doesn’t get named because the AI likes them better. They get named because the AI can verify them more easily.
This is why a firm can rank beautifully on Google and be invisible inside ChatGPT. The two systems are looking for different things.
What actually makes a firm citable
Five inputs determine whether AI tools name your firm. They’re not the inputs most legal marketing budgets are funding right now.
The substance of your reviews matters more than the count. Most firms have spent years optimizing for star ratings and review volume. AI tools don’t care about either of those in isolation. They read the actual text of reviews and pull patterns. A firm with 600 reviews that all read “great experience, highly recommend” is invisible to “best truck accident lawyer” queries because nothing in those reviews tells the AI what kind of cases the firm handles or what specific outcomes it produces. A firm with 200 reviews that describe the type of accident, the insurance dispute, the attorney’s specific work on the case, and the resolution becomes the firm AI cites for those exact intents. The shift in 2026 is from review collection to review substance. You can coach intake teams to ask clients to describe what happened and what got resolved, without scripting the content of the review itself. Google’s current review guidelines prohibit asking customers to include specific content in their review text, including content that names a staff member, so the old “please mention how Sarah helped with your case” template is now a policy violation.
Editorial validation from sources AI systems already trust. When Gemini decides which firm to recommend for the best motorcycle accident attorney in Atlanta, it’s pulling heavily from third-party editorial sources that have made that judgment publicly. Super Lawyers selections. Best Lawyers in America inclusions. Local business journal “Top Lawyers” features. Bar association recognitions. Press coverage of cases your firm tried. CLE speaking engagements with online evidence. Every credible third-party source that names your firm in a list of the best becomes training material for the next AI answer. Firms that have invested in real editorial credibility over a decade have an enormous advantage right now. Firms that haven’t are starting at zero on a clock that’s already running.
Practice area depth on your website. A 400-word page on “car accidents” with stock language and a contact button is invisible to AI tools. They pull from substantive content when generating answers, and shallow content does not get pulled. The pages that get cited are the ones that go deep on the questions clients actually ask before they call. How comparative negligence works in your state. How insurance adjusters value rear-end collision claims. What the statute of limitations looks like for different injury types. What mistakes claimants make in the first 48 hours after a crash. The firms winning citation visibility in 2026 are the ones rebuilding their practice area pages around real depth and writing for the question, not the keyword.
Entity consistency across the web. Your firm name, address, phone, attorney names, practice areas, and core descriptors need to match across your website, your Google Business Profile, Justia, Avvo, Martindale, your state and county bar directories, and the legal directories where you appear. When these signals conflict, AI systems lose confidence and reach for a cleaner entity instead. Most firms have no idea how fragmented their digital footprint is. Different addresses on different directories. Attorney names spelled differently. Phone numbers that no longer ring through to your intake line. Each inconsistency is a small reason the AI picks someone else.
Structured data that connects your entity to the rest of your web presence. Your website’s schema markup is what tells search engines and AI tools what kind of business you are, who your attorneys are, and how your firm’s various web profiles connect into a single coherent entity. Most law firm websites have either no schema or generic schema that establishes nothing. The firms with proper Attorney, LegalService, and LocalBusiness schema with sameAs properties pointing to their authoritative profiles are giving AI tools a machine-readable map of who they are. The firms without it are leaving the AI to guess, and the AI does not guess in their favor.
What good looks like versus what most firms have
Take two personal injury firms in the same mid-sized market. Firm A has 850 reviews averaging 4.9 stars, ranks first or second for every major keyword, has clean GBP optimization, and a website with 30 location pages. Firm B has 220 reviews, ranks fourth or fifth, has a single primary office page with deep practice-area content, has been featured in two local business journal “Top Lawyers” issues, has its lead attorney listed in Super Lawyers for six consecutive years, and has Attorney schema on its site connecting back to Justia, Avvo, and Martindale profiles where the entity information matches exactly.
Ask ChatGPT, Gemini, and Perplexity who the best attorneys in that market are. Firm B gets named. Firm A doesn’t. Firm A has been winning the visible game for years. Firm B is winning the game that’s about to matter.
What most firms are doing wrong
Five assumptions are costing legal marketing budgets right now.
The first is that traditional SEO will keep working as it always has. It won’t, and the data isn’t subtle. Local pack visibility has been compressing for two years and the rate is accelerating. Firms running 2022 playbooks of city pages, keyword density, and 3-pack optimization are losing ground every quarter while their dashboards still show stable rankings.
Treating Google Business Profile as the whole strategy is the second. GBP still matters and is not going away. It’s also a fraction of what AI tools draw from when they decide who to cite. A firm that has only invested in GBP is invisible everywhere else.
Third is reactive panic. A managing partner asks ChatGPT about their firm, gets a generic or wrong answer, and demands an immediate fix. There’s no immediate fix. AI systems update their understanding of an entity based on the entire web’s signal about that entity, and shifting that signal takes months of disciplined work.
Fourth is the persistence of templated location pages. The same paragraph rotated across 30 city URLs with the city name swapped in. AI tools see straight through this and either ignore the pages or downgrade the entity for thin, duplicative content. The architecture that worked in 2018 is now actively damaging.
Fifth is the most expensive to recover from. Vendors are already selling “AI optimization” as a productized service built on prompt manipulation, content tricks, or quick schema kits. Nothing about citation authority is fast. The firms paying for shortcuts in 2026 are paying real money for outcomes that won’t compound.
The work compounds, and the runway is closing
Building citation authority takes six to twelve months of disciplined work to move the needle meaningfully. There’s no version of this where you start in October and dominate by Christmas.
The reason it’s worth doing anyway is that once a firm is established as a recognized entity that AI tools confidently cite, the position is hard to displace. Becoming the default answer captures a disproportionate share of inquiries. More inquiries produce more reviews, more case results, more press, and more authority signals, which strengthens the citation profile next quarter and the quarter after that. Late movers in the prior shifts in legal marketing generally paid more for less ground than the firms who moved early.
The same dynamic is now running in AI search. The firms that started entity authority work in 2025 are positioned to pull away from competitors who haven’t started. The firms that begin in late 2026 will be playing catch-up against a moat their competitors have already built.
What to do next
Run the test. ChatGPT, Gemini, Perplexity. The best [practice area] attorneys in your city. Then ask each one what it knows about your firm specifically. The answers will tell you more about your real market position in 2026 than any ranking report on your desk right now.
If what you find concerns you, the AI Citation Audit will show you exactly which gaps are responsible and what it would take to close them. We run a structured AI Citation Audit for law firms covering four areas: entity consistency analysis across major legal directories and your GBP, schema and structured data evaluation on your website, editorial validation review against your top three competitors, and live citation testing across ChatGPT, Gemini, and Perplexity for the queries that actually matter to your practice. You walk out of it with a ranked list of the specific gaps costing you visibility and a clear sequence for closing them.
Most firms we audit have three to five fixable problems accounting for the majority of their citation invisibility. We can usually identify them within two weeks of engaging us, not two months.
If you want to know what yours are, schedule a 45-minute audit conversation here: Schedule today. The findings will be specific to your firm and your market.
