Client Communication and Marketing
This Firm Function covers everything the firm says to clients and prospects outside the advice itself: marketing copy, social posts, client newsletters, website text, drafted emails, a chat widget answering questions in the firm’s voice, and any AI-generated summary of client testimonials or online reviews. AI reaches this work earlier and more completely than it reaches advice, because nearly all of it is writing, and writing is what these tools do most readily. The constraint is that the rules written about advertisements do not care which tool produced the words. Much of this page describes the SEC’s Marketing Rule, which by its own terms applies to SEC-registered advisers; a state-registered adviser is under the state’s own advertising requirements instead, which this guide does not cover.
Key ideas
The rule attaches to what is said, not to what wrote it
The SEC’s Marketing Rule governs advertisements by investment advisers, and nothing in it turns on how the content was produced. An AI-drafted social post making a performance claim is regulated exactly like one an adviser typed from scratch. There is no lighter standard for AI-drafted marketing and no argument that a tool’s involvement changes what the firm is responsible for having said.
Testimonials and endorsements keep their conditions
The Marketing Rule permits testimonials and endorsements, but only under specific disclosure and disqualification conditions covering things like compensation, material conflicts, and a written agreement with the person giving it. An AI tool that drafts a testimonial, synthesizes several client comments into one quote, or scrapes and rephrases online reviews into marketing copy is producing exactly the kind of content those conditions were written for. The conditions do not relax because a model assembled the words. A quote stitched together from several clients’ comments is not a statement any client made, which the rule’s definition of a testimonial requires, so presenting it as one raises the general prohibitions before the testimonial conditions arise. If anything, this is the use most likely to produce a problem quietly: a tool asked to “write some social proof from our reviews” will produce something that reads like a testimonial without anyone deciding to publish one.
Fair and balanced is where AI-drafted copy fails
The Marketing Rule requires that an advertisement not be misleading, including through omission — that material risks and limitations get fair treatment alongside results. AI-drafted marketing copy has a specific failure mode here. A model asked to write persuasive copy about a strategy’s performance will foreground the favorable framing unless it is explicitly told not to, because that is what persuasive writing means. That instinct runs directly against what the rule requires. Copy produced this way needs the same fact-and-framing review that advice work gets, not a proofread for typos and tone.
A wrong claim in marketing gets repeated at scale
A hallucination in marketing copy looks like a fabricated average return, an invented client outcome, or an overstated years-in-business claim — confident, specific, and wrong. What makes this different from the same error in advice work is reuse. A wrong number in a plan goes to one client, who may well catch it. A wrong number in a one-pager goes to every prospect who receives it, sits on the website for a year, and gets pasted into three other documents before anyone notices. The error is the same; the blast radius is not.
A first-pass checker narrows the review; it does not replace it
The failure modes above are predictable enough to write down once and have an AI tool check every draft against them. One worked example is an open-source checker built for small advisory firms: it reviews a draft against the Marketing Rule, names the paragraph of the rule behind every flag, proposes the smallest fix, and notes where a state’s rules may differ. Its output is a list of questions for the reviewer, not an approval. The person who signs off still owns the review and still decides what the firm says. What changes is where that review starts: with specific, cited flags instead of a blank read.
The chat widget is not a one-to-one conversation
A firm-voice chatbot, or a scripted AI reply tool answering prospect messages, looks like a private exchange with one person. Under the Marketing Rule’s first test, a message offering the firm’s services to prospects is an advertisement when it reaches more than one person, and the adopting release counts bulk or algorithm-generated messages, sent one at a time, as reaching everyone who gets them. A widget giving every visitor the same answers qualifies, even though deploying it feels like installing software, not writing an ad. The release also treats a stored performance figure reused across replies as an advertisement, so a chatbot answering “what returns do your clients get?” from a canned number has made a performance claim.
Marketing materials are records
Marketing materials fall under the same retention obligation as any other advertisement, and an AI-drafted one is still an advertisement. Whether the prompts and intermediate drafts behind it are records too is less settled, and turns on what they contain. See Records and Documentation for what the obligation covers, and this year’s figures for how long things have to be kept. The practical consequence for this area is that the working material matters, not just the published post — if a prompt produced a claim the firm later has to explain, the prompt is part of the explanation.
Where it shows up in the process
- Map the Rules That Apply: The marketing and advertising duty in that step’s table is the one this Firm Function turns on, and any use producing outward-facing content maps to it.
- Handle Client-Facing Use: The communication category in that step is this Firm Function’s territory, and the disclosure question it raises is a marketing question before it is an AI question.
- Review and Keep Records: Published marketing gets reviewed on a cadence like everything else, and the retention obligation above is what that review checks against.
Common misconceptions
- “The Marketing Rule only applies to formal advertisements, not casual social posts.” What counts as an advertisement is broader than a designed campaign, and a performance claim made casually is still a performance claim.
- “If a person reviews the AI draft before it goes out, the testimonial and endorsement conditions don’t apply.” Human review changes who is accountable for catching a problem; it does not change whether the content is a testimonial. The conditions attach to what the content is.
- “Nobody would take a chatbot’s answer as the firm’s position.” A prospect reading the firm’s website has no way to distinguish the firm’s voice from the widget’s, and no reason to try.
- “Marketing copy is lower stakes than advice, so it needs less review.” The stakes are different, not lower — marketing reaches more people, persists longer, and is exactly what an examiner can pull up without asking the firm for anything.
At a larger firm, AI-drafted public content almost always runs through a marketing or compliance pre-approval workflow rather than an individual adviser’s judgment about their own posts. That changes the job from “review my own draft carefully” to “route it through the existing approval queue and wait,” including for content that feels too small to bother with — a LinkedIn post, a reply to a review. The queue exists precisely because scattered individual judgment is what produces inconsistent public claims across a firm. An adviser who finds the approval step slow for AI-drafted content is usually discovering that the firm’s volume expectations and its approval capacity were set before drafting got this fast.
Go deeper
- Investment adviser marketing, 17 CFR 275.206(4)-1 , eCFR — the current text of the Marketing Rule, including the general prohibitions and the testimonial and endorsement conditions.
- Investment Adviser Marketing, SEC Release IA-5653 (2020) — the adopting release, which explains the Commission’s reasoning on what counts as an advertisement and how the testimonial conditions are meant to work.
- Risk Alert: Initial Observations Regarding Advisers Act Marketing Rule Compliance (2024) , SEC Division of Examinations — what examiners actually found when they looked at Marketing Rule compliance, including deficiencies in substantiation and fair-and-balanced treatment.