What do law firms ask about AI visibility?

In the US, yes: since 31 August 2026. OpenAI’s June 2026 ad launch excluded legal services, then opened the category to US advertisers on 31 August 2026; outside the US the ban still stands. But an ad rents a slot next to the answer: it never becomes the answer. A firm that wants to be the recommendation, not the interruption, still needs organic citation.

Make your firm the easiest credible source to cite: consistent name-address-practice data across the web, LegalService and Attorney schema markup, a question-and-answer content hub that directly answers client queries, and crawlable pages open to GPTBot. ChatGPT recommends firms it can identify, verify, and quote: not firms with the biggest ad budget.

We do not quote a percentage here, because we have not found a source we can cite for one. What we can say: legal consumers now start attorney research in AI assistants like ChatGPT, and adoption skews toward exactly the matters consumer firms want: injury, family, and immigration questions people are embarrassed to ask a human first. The audit shows whether those assistants name your firm.

AEO is the practice of making a law firm citable by AI answer engines, ChatGPT, Claude, Perplexity, and Google AI Overviews, so the firm is named when consumers ask legal questions. It combines entity cleanup, schema markup, AI-crawler access, and question-shaped content, and it is measured in citation share-of-voice rather than rankings.

At FirmCited: a USD 490 diagnostic audit, a USD 6,500 Starter install (AEO foundation, 14 days), a USD 12,000 Pro install adding a 24/7 AI intake chatbot (21 days), and a USD 2,000/month growth retainer for ongoing content and monitoring. The USD 490 audit credits in full toward any install within 60 days.

No. Nobody controls what ChatGPT, Claude, Perplexity, or Google AI Overviews say: answers vary by model version, phrasing, and day. Any vendor guaranteeing AI citations or rankings is selling something they cannot deliver. Honest vendors commit, in writing, to what they control: deliverables, timelines, response speed, and transparent measurement.

Typically 6 to 16 weeks from fixing the foundations. Perplexity and Google AI Overviews respond fastest, sometimes within weeks of indexing, because they retrieve live web results. ChatGPT and Claude move slower, as citations depend on crawl cycles and model refreshes. No honest vendor promises a specific date.

An AI visibility audit measures whether AI assistants recommend your firm. FirmCited runs 25 realistic client prompts on up to 4 engines (ChatGPT, Claude, Perplexity and Google AI Overviews) and delivers the report when at least 2 answer. It records every firm cited, computes competitor share of voice, and delivers a 10-page prioritized fix list within 48 hours for USD 490, credited toward any install.

Four: ChatGPT (largest consumer reach), Google AI Overviews (sits atop existing search volume), Perplexity (fast-growing research assistant with visible citations), and Claude (rising consumer use, quality-skewed audience). A firm optimizing for these four covers the overwhelming majority of AI-assisted legal research in the US today.

AI Overviews increasingly answers “lawyer near me” queries with a generated summary naming two or three firms before any traditional result appears. It draws on Google’s index, Business Profiles, reviews, and schema markup. Firms with consistent local entity data and question-shaped content get named; everyone else is pushed below the fold.

llms.txt is a plain-text file at your site root that gives AI systems a concise, structured summary of who you are, what you do, and which pages matter. It costs minutes to ship, carries zero risk, and several AI crawlers already read it: so for a law firm it is a cheap, sensible default.

Four core types: Organization or LegalService for the firm entity, Attorney for each lawyer, FAQPage on question-and-answer content, and Article on substantive pages with a named author and dates. Schema is how AI engines read your identity unambiguously: it converts marketing copy into machine-verifiable facts.

Under 60 seconds. Legal consumers contact multiple firms and typically retain the first one that responds substantively. Response within a minute can multiply qualification rates several times over versus responding in hours: and most firms take four hours or more. Speed-to-lead is the cheapest conversion lever a firm owns.

Yes: when properly fenced. Ethics rules permit technology-assisted intake as long as the bot never gives legal advice, never forms an attorney-client relationship, discloses it is not a lawyer, and keeps confidences secure. The ethical risk is not the chatbot; it is an unfenced chatbot improvising legal conclusions.

No: and a law-firm chatbot must be engineered so it cannot. Applying law to a specific person’s facts is legal advice; delivered by software it risks unauthorized practice, malpractice exposure, and a relationship no one intended to form. Properly built bots inform, qualify, and schedule: then hand facts-specific questions to a lawyer.

AI share-of-voice (SoV) is the percentage of AI-generated answers that cite your firm across a fixed panel of realistic client queries. Run 25 prompts across four engines: if your firm appears in 12 of 100 engine-answers and a competitor in 30, your SoV is 12% against their 30%. It is the core KPI of AI visibility.

Almost always: their web footprint is machine-readable and yours is not. They have consistent entity data across directories, schema markup, crawlable pages, and content that answers client questions directly: so models can identify, verify, and quote them. It is rarely about firm size or reputation; it is about legibility to machines.

No. Paid search ads have zero influence on whether ChatGPT, Claude, Perplexity, or Google AI Overviews cite your firm: AI answers are generated from organic signals, and OpenAI banned legal ads outright. Ads still buy clicks on Google, but every ad dollar leaves your AI-channel invisibility exactly where it was.

Content that answers a real question directly, early, and verifiably: a question-form heading, a complete 40-60 word answer up front, supporting specifics (numbers, statutes, dates, named tools), a table, schema markup, a named author, and a visible update date. LLMs quote sources that make extraction effortless: not the longest or prettiest page.

Track three linked numbers: citation share-of-voice on your money queries (the leading indicator), AI-referred leads, from ask-how-they-found-you intake plus assistant referral traffic (the middle), and retained cases from that channel against average fee value (the result). One retained consumer case typically repays an entire year of AEO investment.

A firm shows up in ChatGPT when the model can identify it as a distinct entity, trusts the sources that mention it, and finds quotable answers on its site. There is no citation to buy: ads rent the slot beside the answer, never a seat inside it. Firms with consistent data, schema markup, and question-shaped content get named; the rest stay invisible.

AEO makes a firm citable by AI answer engines, ChatGPT, Perplexity, Claude, and Google AI Overviews, so it is named when consumers ask legal questions. For a firm it is usually worth it: the channel where legal consumers now start their research has no ad option, and one retained case typically repays a year of work.

Personal injury is the most AI-researched legal category, so visibility matters most there. Queries like “best car accident lawyer in Phoenix” return one to three named firms. PI firms win by combining consistent entity data, pages that answer case-value and fee questions directly, and intake that responds in under a minute.

Market pricing runs from a few hundred dollars for a diagnostic audit, to roughly USD 5,000 to 15,000 for a one-time install, to USD 1,500 to 3,000 per month for an ongoing retainer. At FirmCited: a USD 490 audit, a USD 6,500 Starter install, a USD 12,000 Pro install with an AI intake chatbot, and a USD 2,000/month growth retainer.

Only organically. US law firms can buy ChatGPT ads since 31 August 2026 (elsewhere the category is still banned), but a sponsored slot sits beside the answer: it is never the answer. The way to be cited is being the source ChatGPT selects when it answers a legal question. That makes entity data, schema, and quotable content the entire strategy.

Fix a panel of realistic client prompts, run them on ChatGPT, Claude, Perplexity and Google AI Overviews on a monthly schedule, and count how often your firm is named among the answers that come back. Your citations divided by total answers is your share of voice, the core KPI. Track the monthly trend, because single snapshots are volatile.

Google returns a page of ranked links and ads; a consumer scans several and clicks. ChatGPT returns one generated answer that names only one to three firms, with no legal ads available. For firms, Google rewards ranking and ad budget, while ChatGPT rewards being the single most citable source.

It is a fenced, 24/7 assistant on a firm’s website that greets visitors, answers general questions, qualifies the matter, and books a consultation: responding in under 60 seconds and syncing to the CRM. Properly built, it never gives legal advice or forms an attorney-client relationship; it captures leads that forms and voicemail lose.

With automation that has guardrails. A fenced AI intake agent acknowledges the lead, references what they submitted, asks a qualifying question, and offers a booking, instantly, 24/7, while routing anything substantive to a human. Manual response cannot reliably hit one minute after hours; automation is the only way to hit it every time.

Choose the AEO partner that measures citations transparently, never guarantees them, delivers on stated timelines, and focuses on law firms rather than every industry. Judge vendors by what they promise and how they report: honest ones show raw, prompt-level share-of-voice; the rest promise rankings no vendor can control.

Yes. AI engines cite the most legible source, not the largest firm. A small firm with consistent entity data, schema markup, crawler access, and quotable content is cited over a big firm that is invisible to machines. Size helps only when expressed in signals a model can read: otherwise it is irrelevant.

Four core types, in priority order: Organization or LegalService for the firm entity, FAQPage on question-and-answer pages, Attorney on each lawyer’s bio, and Article on substantive content with a named author and dates. Schema converts marketing copy into machine-verifiable facts, which is how AI engines read your identity without guessing.

For most firms, yes: as a cheap, low-risk default. llms.txt is a plain-text file at your site root that gives AI systems a concise summary of who you are and which pages matter. It takes minutes to ship, several AI crawlers already read it, and it reduces the entity confusion that keeps firms invisible.

Perplexity retrieves live web results and shows its sources, so it favors fresh, indexed, schema-marked pages that answer a query directly. Publish quotable question-shaped pages, keep them crawlable and current, ensure consistent entity data, and Perplexity often cites within weeks: faster than ChatGPT, whose citations follow slower crawl and refresh cycles.

The blueprint is identical across practice areas, entity data, schema, crawler access, and quotable answers, but the queries and nuances differ. Family and estate consumers research emotionally charged decisions privately; immigration consumers often ask in other languages. Each area wins with dedicated, schema-marked pages answering its specific money queries.

GEO, generative engine optimization, is the practice of making a law firm the source an AI engine chooses to quote when it answers a legal consumer’s question. Where SEO optimizes for ranked links, GEO optimizes for being named inside a single generated answer on ChatGPT, Perplexity, Google AI Overviews, or Claude.

The same way they get clients from Google, by being the firm the search result names, except the result is a single generated answer, not a list of links. A firm gets more AI-search clients by being cited: consistent entity data, schema-marked direct answers to client questions, and quotable content across the engines consumers actually use.

Make it crawlable, identifiable, and quotable. ChatGPT cites firms it can read and trust: allow GPTBot in robots.txt, keep firm name-address-practice data consistent, add LegalService and FAQPage schema, and publish direct question-and-answer pages that answer client queries in the first sentence.

The cost is the clients who research in AI and never see the firm. The firms not cited in those answers lose the matter before the consumer ever visits their site.

A law firm's digital footprint is every place the firm appears online: its website, Google Business Profile, Avvo, Justia, LinkedIn, press, and directories. AI engines use that footprint to decide the firm exists, what it does, and whether to trust it. Inconsistent footprints get firms dropped from answers; consistent ones get them cited.