SEC Marketing Rule checklist: pre-reviewing advisor marketing with an AI first pass

What counts as an advertisement

The Marketing Rule, Rule 206(4)-1 under the Advisers Act, has applied since November 4, 2022. Its definition of an advertisement has two parts. The first covers communications to more than one person, or to one person when hypothetical performance is included, that offer advisory services to prospective clients or new advisory services to current clients. It excludes some communications, such as extemporaneous, live, oral ones. The second part covers testimonials and endorsements the adviser pays for, directly or indirectly. Have the CCO classify each piece under paragraph (e)(1). Newsletters, social posts, seminar decks and website pages usually fall inside the definition, while a one-to-one recap email after a review generally falls outside it and is still a record under Rule 204-2.

This guide is general information about an SEC rule, not legal or compliance advice. State-registered advisers should check their state's rules. For communications subject to FINRA Rule 2210, which covers a member firm's communications with the public, a registered principal generally has to approve retail communications before use, subject to the rule's exceptions.

Examiners keep finding problems. The SEC's Division of Examinations published Marketing Rule risk alerts in April 2024 and December 2025. A September 2026 alert on annual compliance reviews also found marketing policies and procedures that had not been updated for the rule. In a small firm the CCO is often also the COO or a principal, reviewing every post personally, and an AI first pass can save that person real time.

The pre-review checklist

Run every piece through these checks before use. Each one ties to a part of the rule, so the reviewer can cite it when sending a piece back.

• Truth and omissions, paragraph (a)(1): no untrue statement of a material fact, and nothing left out that makes the rest misleading.
• Substantiation, paragraph (a)(2): a reasonable basis to believe the firm can substantiate every material statement of fact if the SEC asks.
• Implications, paragraph (a)(3): no wording likely to lead a reader to an untrue or misleading conclusion about a material fact.
• Benefits and risks, paragraph (a)(4): any potential benefit comes with fair and balanced treatment of its material risks or limitations.
• Specific advice, paragraph (a)(5): any reference to past recommendations is presented in a fair and balanced way.
• Performance, paragraph (a)(6): periods and results are chosen and shown fairly, and the performance conditions in paragraph (d) are met.
• Hypothetical performance, paragraph (d)(6): relevant to the intended audience, with its assumptions and limitations explained.
• Testimonials and endorsements, paragraph (b): clear and prominent disclosures, a reasonable basis that each one complies, and the promoter checks described below.
• Third-party ratings, paragraph (c): the questionnaire behind the rating makes favorable and unfavorable answers equally easy, and the required disclosures appear clearly and prominently.
• Everything else, paragraph (a)(7): nothing in the piece is otherwise materially misleading.

Performance, promoters and ratings in more detail

Performance has its own rules in paragraph (d). Gross performance needs net performance beside it with at least equal prominence. Portfolio performance generally needs 1, 5 and 10-year periods, and no piece may say or imply that the SEC approved or reviewed the calculation. Hypothetical performance needs policies that make it relevant to the intended audience's likely financial situation and objectives, plus enough information for that audience to understand the assumptions used and the risks and limitations of relying on it.

Testimonials and endorsements need the most care. A testimonial is a statement by a current client about their experience with the adviser, and an endorsement comes from someone who is not a current client. Both need clear and prominent disclosure of whether the person is a client and whether they were paid, plus a brief statement of any material conflict. Check promoter eligibility too, because the rule bars paying a promoter the firm knows, or should know, is disqualified. A paid promoter also needs a written agreement. Both checks fall away when the pay is de minimis, which the rule sets at $1,000 or less over the preceding twelve months, and the firm's own staff have a separate exemption when their affiliation is apparent or disclosed.

Third-party ratings need two checks. The questionnaire or survey behind the rating has to make favorable and unfavorable answers equally easy, and it must not be designed to produce a set result. The piece must also disclose the date of the rating and the period it covers, along with who created it and any payment by the firm connected to it.

The mistakes SEC examiners have written up

The two Marketing Rule risk alerts are the best test cases available, because they describe what examiners actually found. The December 2025 alert focused on testimonials and endorsements. Examiners found required disclosures missing or tucked behind hyperlinks, and disclosures set in smaller or lighter font than the testimonial. They also found advisers that copied current client testimonials and former client endorsements from third-party websites onto their own sites without clearly disclosing that clients had written them. If your firm shares client reviews from Google or similar sites, read that alert and the staff FAQ before you post another one.

Claims about AI get their own scrutiny. In March 2024 the SEC charged two advisers, Delphia and Global Predictions, with false and misleading statements about their use of AI, and both settled for $400,000 in combined penalties. Global Predictions also violated the Marketing Rule. The FY2026 examination priorities say examiners will review registrant representations about AI capabilities for accuracy. A firm that tells clients it uses AI has to describe that use accurately, and we hold our own pages to the same standard.

The earlier April 2024 alert lists these findings, among others:

• Advertisements stating the adviser was "free of all conflicts" when conflicts existed.
• Statements that a team served clients when one individual did the work, and errors in staff qualifications such as designations.
• References to investment mandates or screening processes the firm did not have.
• Publicity for awards and accolades the firm had not actually received.
• Calling "best interest of clients" a differentiator without disclosing that every adviser owes clients that duty.
• Claims of being "seen on" national media when the appearances were paid advertisements, and celebrity images implying an endorsement that did not exist.
• Net performance calculated with lower fees than the intended audience would pay.
• Claims to consider risk tolerance when every client was placed in the same strategy.

What the agent flags and what the CCO decides

An AI first pass reads each draft against the checklist and returns flags, each tied to a sentence and a rule paragraph, with a request for evidence behind every material claim. The CCO decides each flag. The agent never approves a piece or publishes it.

FINRA's guidance points the same way for dually registered firms. Regulatory Notice 24-09 says FINRA's rules are technology neutral and that Rule 2210's content standards apply whether a human or a technology tool generated the communication. Where Rule 2210 applies, the principal approval it generally requires for retail communications does not change because an AI drafted the piece.

A first pass is useful only when the CCO knows what it misses. Test it on pieces you have already reviewed, including ones you sent back, and check that it catches what you caught. Keep sampling its output every month after that, and keep rewriting with the marketing team, since the agent suggests what needs evidence or balance and leaves the wording to people.

The records file for each piece

Rule 204-2 asks for a copy of each advertisement, records of the testimonial and endorsement disclosures, the workpapers behind any performance shown, and a record of the intended audience for hypothetical performance. Advertisements and performance records are kept for five years from the end of the fiscal year in which the piece was last published or disseminated, the first two years in an appropriate office of the adviser.

Last use is the date that matters. A website page that stays up for three years runs its retention clock from the end of the fiscal year in which it came down. The agent can build that file as it works: the approved version, each flag with the CCO's decision, the evidence for every claim, the intended audience and the dates of each use.

Social posts and texts need archiving as well. In the 2026 T3 survey, 40.76% of respondents used a social media and text message archiving tool, led by MyRepChat and Smarsh. Whatever you use, the approved version in the archive should match the version the CCO signed.

Worked example: pre-review hours for a busy CCO

This is an illustrative example with assumed numbers, not a client result. Take a firm that produces 20 pieces a week across advisor social posts, a newsletter, website edits and a seminar deck. The CCO spends 25 minutes on each today, so 500 minutes, or about 8.3 hours a week.

Assume 15 of the 20 are routine, with no performance figures or client testimonials. With an agent's first pass, the CCO spends about 8 minutes on each, 120 minutes in all. The 5 pieces that show performance or client reviews still take the full 25 minutes, another 125 minutes. Total time falls to 245 minutes, about 4.1 hours, which gives back about 4.3 hours a week.

On its own that does not reach the sprint's 10-hour guarantee, and we would say so on the call. The bigger gain is the file: each approved piece is stored with its flags, evidence, audience and last-use date, ready for an examiner. Marketing review fits best in a sprint alongside meeting follow-up or account paperwork.

Starting small with marketing review

Collect everything your firm published last quarter and run it through the checklist yourself, noting the minutes per piece and what you flagged. Read your written marketing policies against the rule, since the September 2026 risk alert found firms that had not updated theirs. Then test an AI first pass on the same pieces and compare its flags with yours.

Our five-day AI Workflow Sprint can run that comparison on your own material and price a build. It costs $2,500, with a full refund if we do not find 10 staff hours a week. Whatever tool you choose, the CCO's name stays on every approval.

Common questions