What an AI-visibility vendor can legally promise your law firm

<!-- PUBLISHER: INBOUND LINKS STILL OWED TO THIS PAGE (design/content/CONTENT_PLAN.md, B2). These live on pages outside this file and must be wired at publish time: 1. /methodology → link to /blog/what-ai-visibility-vendors-can-promise-law-firms (REQUIRED by CONTENT_PLAN.md B2, "Inbound: link from /methodology and from B6".) 2. B6 (the vendor-diligence piece, "is AI visibility a scam") → link here once, when B6 ships. B2 and B6 each link to the other exactly once. Also on ship: the report-language table below is the designated linkable asset — keep its anchor id and the attribution line beneath it stable across revisions. -->

There is a sentence in the proposal on your desk. It reads something like "We'll get your firm recommended by ChatGPT" — or, in bolder versions, "guaranteed placement in AI answers." Before you price the engagement, price the sentence. If it reaches your marketing, you answer for it.

Stated plainly first: Briefly is not a law firm, holds no legal credentials, and this is not legal advice. What follows reports what published rules say and links to the text. Whether a particular sentence complies in your jurisdiction is for your ethics counsel or your state bar to say.

Who carries the exposure

The vendor writes the sentence. The firm answers for it. Three short provisions do the work.

Rule 8.4(a) makes it misconduct for a lawyer to "violate or attempt to violate the Rules of Professional Conduct, knowingly assist or induce another to do so, or do so through the acts of another" (ABA Model Rule 8.4). An agency is another.

Rule 5.3 extends supervisory duty to nonlawyers, and its Comment reaches past the payroll. Comment [1] covers "nonlawyers in the firm and nonlawyers outside the firm who work on firm matters." Comment [3] sets the standard for those outside vendors: "a lawyer must make reasonable efforts to ensure that the services are provided in a manner that is compatible with the lawyer's professional obligations" (Comment on Rule 5.3).

One caveat, because this is the only place on the page where we read a rule rather than report it. Comment [3]'s examples are vendors who help render legal services: an investigative service, document management, printing, cloud storage. A marketing agency is not on that list. Comment [1]'s "work on firm matters" is written more broadly, and we think it reaches an agency speaking in the firm's name. That is our reading, not a ruling.

Rule 7.2(d) requires any communication made under that rule to include "the name and contact information of at least one lawyer or law firm responsible for its content" (ABA Model Rule 7.2). The rule asks for a name, and it is not the vendor's.

The vendor is not exposure-free — it answers to a different regulator. The FTC's September 2024 Operation AI Comply sweep included a seller of "AI Lawyer" services; the February 2025 order imposed $193,000 in relief. Two regimes, two defendants. The bar's version lands on you.

What Rule 7.1 actually says

The rule the vendor's sentence has to survive is one sentence long:

"A lawyer shall not make a false or misleading communication about the lawyer or the lawyer's services. A communication is false or misleading if it contains a material misrepresentation of fact or law, or omits a fact necessary to make the statement considered as a whole not materially misleading." (ABA Model Rule 7.1)

Its reach is settled in the first line of its own Comment: "This Rule governs all communications about a lawyer's services, including advertising." Any statement about what your firm is or does, wherever it appears and whoever typed it, is inside the rule.

The operative word arrives in Comment [3], and it is unsubstantiated:

"an unsubstantiated claim about a lawyer's or law firm's services or fees, or an unsubstantiated comparison … with those of other lawyers or law firms, may be misleading if presented with such specificity as would lead a reasonable person to conclude that the comparison or claim can be substantiated."

That is more permissive than most firms assume and more demanding than most vendors admit. The rule does not forbid specificity. It forbids specificity you cannot back.

The test is not new, and no new opinion is needed to supply it. Advertising conduct is governed by the 7.x rules, and Rule 7.1 asks one question of every sentence a vendor hands you: can you substantiate it?

States are not the Model Rules. Every jurisdiction writes its own version, and advertising is among the provisions states rewrite most — lettering included, so your Rule 7.2(d) may carry a different letter. The ABA maintains jurisdictional rules comparison charts. Read your state's text first.

The AI ethics opinions, and the one that reaches advertising

Two documents get cited in almost every vendor deck. Both are about a lawyer using AI to do legal work.

ABA Formal Opinion 512 (July 29, 2024) was the ABA's first ethics guidance on lawyers' use of generative AI. americanbar.org blocks automated retrieval, so we could not open the PDF; we describe it only as published summaries do. The Bar Examiner, the National Conference of Bar Examiners' journal, put its scope this way: "The ethical issues discussed relate to lawyers' duties of competence, confidentiality, communication, supervision, candor, and duty to charge reasonable fees." The Florida Bar News described it the same way on July 31, 2024.

Texas Opinion 705 (February 2025) takes the same posture: "lawyers are responsible for the work product they submit regardless of who (or what) does the original research and drafting." Work product, not marketing.

One state opinion reaches advertising directly. The Florida Bar's Ethics Opinion 24-1 (January 19, 2024) applies Florida's advertising rules to AI:

"Lawyers may advertise their use of generative AI but cannot claim their generative AI is superior to those used by other lawyers or law firms unless the lawyer's claims are objectively verifiable."

Florida binds nobody else, but objectively verifiable is a useful test to hold a draft against. The opinion adds that "whether a particular claim is capable of objective verification is a factual question that must be made on a case-by-case basis" — so nothing below is a ruling on your wording.

The promise ladder

Four rungs. What matters is not how modest the language sounds, but whether the sentence predicts an outcome or describes a procedure.

Only rung 4 survives the "objectively verifiable" test, for an unglamorous reason: every clause is a fact about what the vendor did, not a prediction about what a model will do. It is also the only rung that yields a defined number, such as Case Recommendation Share. Ask a vendor at rungs 1 to 3 for the method behind the number they intend to report. No methodology page with prompts, models, date windows and sample sizes means the number is not a measurement.

What a firm can publish about its own AI results

Every firm, figure, and date below is invented, to show the shape of a defensible sentence. Swap in your own measured numbers; do not publish these.

Reuse this table. Copy it into your own guidelines or compliance page, and please include this line:

Source: Briefly, "What an AI-visibility vendor can legally promise your law firm," https://www.askbriefly.ai/blog/what-ai-visibility-vendors-can-promise-law-firms

The left column shares three properties: a stated date window, a stated method, and no comparison to firms nobody measured. The right shares one — it invites a reasonable person to conclude the claim can be substantiated, which is the Comment's own language.

Why no vendor can promise the outcome

A vendor cannot substantiate an outcome promise, because the underlying system does not behave like one.

• Answers are probabilistic. A 2026 audit of retrieval-augmented commercial recommendation found identical reruns of one prompt produced recommendation sets overlapping only ~50–61%, and cosmetic paraphrases dropped that to ~29% (arXiv, 2026 — an unreviewed preprint from an AI-visibility vendor, measured on consumer buying questions rather than legal ones; we cite it against our own interest, since it argues per-prompt tracking is unreliable).
• A model update can reshuffle everything. Assistants retrain on their own schedules with no notice. A firm at the top of every answer in June can be absent in August without changing a word on its site.

So "your firm will be recommended" is a forecast about a system nobody controls, made by a party with an interest in it. The sourced evidence sits on our AI search statistics hub.

What Briefly will not promise you

A rule a vendor exempts itself from is just marketing, so:

Briefly cannot promise your firm will be recommended by any AI assistant. Not first, not often, not at all. No vendor can, and one that does is handing you a sentence you will be the one to defend.

Then the part that costs us something. A Briefly report becomes a Rule 7.1 problem the moment a firm paraphrases it into a superlative. If we report your firm was recommended in 34% of tracked answers in a stated window, and that becomes "AI's most-recommended firm" on your homepage, the measurement was fine and the marketing is not. Keep the number attached to its date window, question set, and method. Drop every comparison to firms we did not measure. Our editorial policy holds us to the same standard.

Briefly measures what AI assistants say. It does not rank, rate, or endorse attorneys, and nothing here is legal advice.

FAQ

Can any vendor guarantee my firm will be recommended by ChatGPT? No. AI assistants generate answers probabilistically and change without notice, so no party outside the model developer can commit to an output. A vendor can commit to the work, and to reporting what happened.

Who gets disciplined if a marketing vendor makes a false claim about my firm? Under the Model Rules, the lawyer. Rule 8.4(a) reaches conduct done "through the acts of another," Rule 5.3's Comment extends supervisory duty to nonlawyers outside the firm who work on firm matters, and Rule 7.2(d) requires a lawyer or firm to be named as responsible for the content.