AEO for law firms: getting named by the engines without crossing Rule 7.1
Last updated October 14, 2026. Written by the Third Coast AI team.
Answer engine optimization for a law firm is the work of getting the firm and its named attorneys recommended when someone asks ChatGPT, Google AI Overviews and AI Mode, Gemini, Perplexity and Copilot which lawyer to call for a specific problem in a specific place, and it is won on practice pages that state what the firm handles in plain, substantiated sentences, on attorney profiles the engines can resolve as real people, and on a record at the State Bar of Michigan directory, Avvo, Google and the local press that agrees with the site. Everything on those pages is a communication about a lawyer's services under Rule 7.1 of the Michigan Rules of Professional Conduct, so the method has to be built around the rule rather than worked around it.
We wrote this page for the firms of West Michigan, where we meet clients in person, and the mechanics and most of the compliance notes apply to a firm anywhere, with the substitution of your own state's version of Rule 7.1. It covers what legal prompts look like, which decision factors come back when we mine them, which third-party sources the engines cite for lawyers, what the rule permits, and a plain plan.
Which questions do people ask the engines about lawyers?
Legal prompts are among the most specific prompts there are, because the person asking has a problem and usually a deadline. In our prompt banks they look like "employment lawyer in Grand Rapids for a non-compete I was just handed," "estate planning attorney in Holland who handles a family cottage going to three kids," "business litigation firm in Detroit for a supplier contract dispute under $500,000," "family law attorney in Kalamazoo who does collaborative divorce," and "do I need a lawyer for a first OWI in Ottawa County, and who." The engine treats each as a shortlist question and names two to five firms or attorneys with a sentence on each.
There is a second kind of prompt that firms underrate, which is the question asked before the person knows they need a lawyer: "is a non-compete enforceable in Michigan," "how does cottage succession work." The engines answer those from pages that explain the law plainly, and the firm whose page is cited for the explanation is often the firm named when the follow-up question is "who should I call." Pew Research measured click-through on Google results pages at 8% when an AI summary was present against 15% without one (July 2025), so the explanation page is read inside the answer far more often than it is visited.
How much of the engine's choice comes down to a fixed set of factors?
| Decision factor | How the prompt phrases it | What counts as proof on the page |
|---|---|---|
| Practice focus | "who actually handles non-competes" | A page per matter type that says what the firm does and does not take, in sentences |
| Named attorneys | "which attorney at the firm does this" | A profile per lawyer with bar admissions, years, matter types, and a P number the directory can confirm |
| Matter types and experience | "has handled cases like mine" | Representative matter types described without unjustified expectations about results |
| Fee structure | "flat fee or hourly" / "free consultation" | The structure stated plainly and kept current |
| Location and courts | "appears in Kent County circuit court" | Office address, counties served, courts where the firm regularly appears |
| Reviews and standing | "well reviewed" / "no disciplinary history" | Google and Avvo profiles that exist and agree with the site; bar standing confirmable in the directory |
| Responsiveness | "will call back today" | Intake process and typical response time stated |
| Languages and accessibility | "speaks Spanish" | Stated on the profile, where true |
Most firm sites state practice areas as a list of nouns and leave the rest for the consultation, which the engine cannot attend. Aggarwal and colleagues at Princeton (KDD 2024) measured what specific, sourced writing does for citation, with statistics lifting a page's visibility in generated answers by about 33%, cited sources by about 28% and quotations by about 41%, and for a law firm the "statistics" are things like years, counties, admissions and fee structures rather than case results.
Do directories matter more than the firm's own site?
- The State Bar of Michigan member directory, which says it allows you to check an attorney's status and locate contact information, searchable by name, P number, county, practice area and language, and which states that membership information does not demonstrate training, experience or expertise. The engines use it to confirm that the named lawyer exists and is in good standing.
- Avvo, described in its Wikipedia entry as an online marketplace for legal services whose lawyer profiles include client reviews, disciplinary actions, peer endorsements and lawyer-submitted legal guides. A claimed, accurate profile is cited; an unclaimed one with a wrong practice area is cited too, with the wrong practice area.
- Google Business Profile and its reviews. Google's local ranking help page says local results are ranked on relevance, distance and prominence and that more reviews and positive ratings can help a business's local ranking, and the AI features read the same profile.
- Bar association and section pages, including local bar associations and the sections a lawyer sits on, which the engines read as corroboration of practice focus.
- Local press and business journals for firm news, hires and commentary, and the firm's own explanatory pages for the law itself.
The 5W citation index attributed 85.5% of AI citations to earned, third-party media rather than brand-owned content, and Ahrefs' 75,000-brand study found brand web mentions correlated with AI Overview visibility at 0.664 against 0.218 for backlinks. For a firm that means the directory and review record matters at least as much as the site, which the reviews and reputation page covers in detail.
Does Rule 7.1 limit what an AEO page can say?
It sets the boundaries, and they are workable. Rule 7.1 of the Michigan Rules of Professional Conduct permits any form of public communication about a lawyer's services that is not false, fraudulent, misleading or deceptive, and says a communication shall not contain a material misrepresentation of fact or law or omit a fact necessary to make the statement not materially misleading, shall not be likely to create an unjustified expectation about results the lawyer can achieve, and shall not compare the lawyer's services with other lawyers' services unless the comparison can be factually substantiated. The comment adds that the prohibition on unjustified expectations would ordinarily preclude advertisements about results obtained for a client, such as damage awards or a record of favorable verdicts, and would ordinarily preclude advertisements containing client endorsements.
In practice that rules out "best employment lawyer in Grand Rapids," because the comparison cannot be substantiated, it rules out case results and verdict records on the practice pages, and it rules out client testimonials pasted into the site as advertising, which is a separate question from clients leaving reviews on third-party platforms in their own words. What the rule leaves is exactly what the engines want: plain statements of what the firm handles, who handles it, where, for what fee structure, and plain explanations of the law. Rule 7.2(b) also requires a copy or recording of each advertisement to be kept for two years after its last use, so every page we publish is archived with its date, and Rule 7.2(c) prohibits giving anything of value for a recommendation, which rules out paid placements in the entity work.
Our compliance process for law firm pages
Every page is reviewed and approved by the responsible attorney before it goes live, every claim about experience is phrased as matter types rather than outcomes, every comparison is removed unless the partner can substantiate it, and the publish gate includes a Rule 7.1 check alongside the factor and sourcing checks. The FTC's Consumer Reviews and Testimonials Rule, announced August 14, 2024, separately prohibits buying reviews or conditioning incentives on positive ones, which we treat as applying to review requests regardless of the bar rules.
The rule and the engine want the same thing, which surprises most partners. Say precisely what you handle, name who does it, do not promise outcomes, and the engine has what it needs to recommend you while the bar has nothing to object to.
Third Coast AI
Can you describe a plain plan for a firm?
- Prompt bank by practice group. 60 to 120 prompts per group, written the way people describe their problem, with city, county and court variants, across visibility, decision-factor, competitor and fact-check prompt types.
- Mining with retrieval on. Each prompt run more than once across the engines, with the hidden searches and the citations captured, so we know which directory, which press and which competitor pages the engine is reading for your matter types.
- Proof audit. The factor list checked page by page against the site and the attorney profiles, with a compliance read of every existing claim.
- Practice and profile pages. One page per matter type and one profile per attorney, written with the partner, stating focus, admissions, counties, courts and fee structure, plus explanatory pages on the law your buyers ask about first.
- Record reconciliation. State Bar directory data, Avvo, Google, bar section pages and LinkedIn brought into agreement with the site, a Wikidata item created for the firm, and a correction page for anything the engines currently get wrong.
- Publish gate. Factor coverage, sourcing, structure and the Rule 7.1 review, with the attorney's sign-off recorded and the page archived for the two-year retention requirement.
- Verification. The prompt set re-run after indexing and monthly after that, tracked in PromptWatch and Profound alongside our own prompt miner, reported at the prompt level with the misses included.
The audit takes about three weeks and usually covers two or three practice groups. The ongoing work runs through the GEO content program and the citation and entity program, fees are quoted after a short scoping call, and firms that want to run the method in-house can read the agencies and professional-services page about training. The Grand Rapids and Lansing pages cover the two Michigan markets with the most firms.
Frequently asked questions
Not in Michigan, unless the comparison can be factually substantiated, which it almost never can. Rule 7.1 prohibits comparisons that cannot be substantiated, and the engines do not need the superlative anyway; they need to know what you handle and who handles it.
The comment to Michigan's Rule 7.1 says the prohibition on unjustified expectations would ordinarily preclude advertising results such as damage awards or a record of favorable verdicts. We describe matter types and experience instead, and the responsible attorney approves every page.
Reviews written by clients in their own words on a third-party platform are a different thing from the firm publishing endorsements as advertising, and we handle review requests under the FTC's rule on reviews, with no incentives conditioned on a positive review. Ask your ethics counsel how your firm wants to treat quoting them on the site.
Both, and often the lawyer first. Attorney profiles that the State Bar directory can confirm and that agree with Avvo and LinkedIn are what let the engine resolve the person as an entity.