Every advisor writes the same quarterly commentary to 200 households. Every advisor also writes — or should write — a fundamentally different concept memo to a 58-year-old executive than to a 71-year-old retiree, even when the underlying planning concept (Roth conversion, NUA, QCD, 72(t), mega-backdoor Roth) is identical. The first job is mass-produced, the second is fully personalized, and both are subject to SEC Rule 206(4)-1 the moment they touch a client's screen. The 2024-2025 AI-washing settlements (Delphia, Global Predictions, the 2025 enforcement cluster), the SEC Division of Examinations Risk Alerts, and the January 2026 staff FAQs on third-party ratings, hypothetical performance, and testimonial mechanics define exactly how AI-generated commentary and concept memos must be drafted, disclosed, substantiated, and archived in 2026. This lesson installs the workflow that produces both the quarterly market commentary and the personalized concept memo from one prompt template — with the forward-looking-statement-safe market commentary balanced under 206(4)-1(d), the household-tailored concept memo carrying documented sources and the "clear and prominent" disclosure for AI-generated content, and both passing the firm's Marketing Rule pre-use review queue without becoming the next AI-washing enforcement headline.

Why This Is the Most-Scrutinized AI Output in 2026

Every word an advisor sends to a client is a written communication. Every written communication that offers advisory services or describes investment results is an "advertisement" under Rule 206(4)-1. Quarterly commentary describes investment results and forward-looking expectations. Concept memos describe investment strategies and recommend household action. Both are advertising. The 2024-2025 AI-washing settlements turned on misrepresentation of AI capability in marketing language; the January 2026 staff FAQs added flexibility on certain third-party-rating mechanics but did not relax the "clear and prominent" disclosure standard or the substantiation requirement; the FINRA 2026 Regulatory Oversight Report's framing of agentic AI under Rule 3110 makes principal review of AI-drafted client content a supervisory expectation rather than a nice-to-have.

The operational consequence: every AI-generated quarterly commentary and every AI-generated concept memo must (a) be substantiated against the firm's actual operation, (b) carry a "clear and prominent" disclosure of any AI-generated content where the firm's posture requires it under the January 2026 FAQs interpretation, (c) survive the firm's pre-use principal review under FINRA Rule 2210 (for hybrid practices) and Marketing Rule substantiation review (for RIAs), (d) avoid hypothetical performance under 206(4)-1(d) unless audience-tailored, (e) avoid cherry-picked performance presentations, and (f) be retained as records under SEC Rule 204-2 and FINRA Rule 4511 — including the prompts, AI tool version, outputs, advisor edits, and reviewer signoff. The substantiation file (built in L4 Ch7) and the pre-use review queue (built in L4 Ch3) are the standing infrastructure this lesson assumes.

Quarterly Commentary — The Mass-Produced, Same-for-Every-Household Artifact

The quarterly commentary is the one piece of content that goes to substantially all households substantially unchanged. It is therefore the highest-leverage place to install AI assistance and simultaneously the highest-risk place for Marketing Rule trips because every household reads the same forward-looking-statement language and any single misrepresentation echoes across the book. The 2026 advisor convention has converged on a five-section structure: (1) the quarter's market summary (US equity, international, fixed income, real assets, alternatives), (2) the macro context (Fed posture, fiscal policy, geopolitical drivers, inflation print sequence), (3) the forward-looking commentary (themes, risks, opportunities — explicitly framed as opinion, not forecast), (4) what this means for the firm's strategic ranges (the IC's tilts within the IPS-defined ranges, never a tactical timing call that conflicts with the IPS), and (5) the standing disclosures.

The Quarterly Commentary Prompt

A representative working prompt for a May 2026 firm: "You are the firm's quarterly market commentary author. Draft the Q2 2026 quarterly commentary in the firm's standard five-section format. Inputs: (a) the quarter's index returns by asset class from [data source — e.g., Morningstar, Bloomberg, the firm's preferred data vendor], (b) the firm's IC-approved macro view document, (c) the firm's standard disclosure boilerplate, (d) prior-quarter commentary for tone consistency. Constraints: (1) Forward-looking language must be framed as opinion or possibility, never as forecast or guarantee. (2) Do not present any specific past performance numbers without the corresponding period-over-period comparison and the index methodology disclosure. (3) Do not include any hypothetical performance, model performance, or 'what if you had invested' content under SEC Rule 206(4)-1(d). (4) Do not include any specific household-identifying examples; the commentary is firmwide. (5) Avoid cherry-picked performance — if a single asset class is highlighted, the equivalent treatment must be applied to the lagging asset classes in the same quarter. (6) Cite every macroeconomic data point with the source and date (Bureau of Labor Statistics CPI release date, Fed FOMC statement date, etc.). (7) End every forward-looking paragraph with the standing 'past performance is not indicative of future results; this is not a recommendation to buy or sell any security' disclaimer language from the firm's boilerplate. Output the commentary in Markdown with section headers and the disclosure block at the end."

The Forward-Looking-Statement-Safe Discipline

The lesson's hardest discipline is framing forward-looking commentary as opinion rather than forecast. The line is thin. "The Fed is likely to pause in Q3" is opinion; "The Fed will pause in Q3, opening a window for fixed-income reentry" is forecast plus implied recommendation, and the latter is a Marketing Rule risk. The AI's strongest failure mode here is confident-sounding forecasting language because it pattern-matches against confident financial writing in its training data. The prompt must explicitly constrain ("frame as opinion or possibility, never forecast"), and the senior advisor's review pass must catch the residual confidence the AI sneaks in. The L1 Ch4 L1 lesson covers the Marketing Rule framing; this lesson applies it to the recurring quarterly artifact.

The Cherry-Picked-Performance Trap

Under 206(4)-1(d), an advertisement that includes performance must present it fairly. A commentary that says "US large-cap returned 8.4% in Q1, demonstrating the strength of our equity exposure" without also acknowledging that international equity returned 2.1% and US small-cap returned -1.3% (and without the firm's actual portfolio exposure breakdown) is cherry-picking. The AI must be prompted to present asset-class performance comparably across all major classes the firm allocates to. The senior advisor's review pass confirms balance and adds any household-portfolio context the commentary needs.

Personalized Concept Memos — Same Topic, Two Different Households, One Prompt Template

The personalized concept memo is the inverse problem: same planning concept, fundamentally different memo content depending on the household's facts. The lesson works two examples in parallel: a 58-year-old technology executive household (the Wilsons — $4.2M, $1.4M unexercised ISOs, $2.1M pre-tax 401(k), high marginal bracket, no current charitable strategy) and a 71-year-old retiree household (the Hendersons from prior lessons — $4.9M post-inheritance, ongoing RMD coordination, recently-expanded charitable orientation). Both households ask "should I be thinking about a Roth conversion?" The two memos differ on every operative dimension.

The Concept Memo Prompt Template

The same prompt template produces both memos: "You are the firm's planning concept memo author. Draft a personalized concept memo for the household on the topic of [CONCEPT]. Inputs: (a) the household's IPS Edition N, (b) the household's most recent Holistiplan extraction of the prior-year 1040 with current-year projection, (c) the household's RightCapital / eMoney / MoneyGuidePro plan, (d) the household's discovery-update transcript or most recent meeting notes, (e) the firm's standard disclosure boilerplate. The memo must (1) explain the concept in plain English the household can read aloud, (2) present the specific decision the household faces with their actual numbers, (3) cite every IRS code section and every IRS publication the analysis depends on (e.g., IRC §408A for Roth IRA rules, IRC §401(a)(9) for RMD coordination, IRC §72(t) for early-withdrawal penalty, IRC §402(e)(4) for NUA, IRC §408(d)(8) for QCDs, IRC §409A for deferred comp), (4) present the alternatives considered with the trade-offs (do nothing, partial action, full action, multi-year sequence), (5) identify the household-specific factors that drive the recommendation, (6) end with a clear next-step recommendation and the senior advisor's role in execution. Do not invent specific dollar figures — extract them from the IPS, Holistiplan, and planning software inputs only. Do not present hypothetical performance. Include the 'clear and prominent' AI-generated-content disclosure if the firm's posture requires it. End with the firm's standard concept-memo disclosure boilerplate. Output in Markdown with section headers."

The Wilsons' Roth Conversion Memo

The 58-year-old executive household. The memo opens with the concept ("Roth conversion is the act of moving pre-tax retirement dollars to a Roth account, paying tax today to avoid tax in retirement"). The memo's analysis specific to the Wilsons: marginal bracket is 35% federal + 3.8% NIIT + 9.3% California state = 48.1% combined; pre-tax 401(k) is $2.1M; current-year unrealized ISO income is paper-only (not yet AMT-triggering); the conversion window of "low-income years" is not currently open because the household is in peak earning years; the conversion would be tax-expensive today. The memo's recommendation: defer the Roth conversion to the post-separation window (when the household exits the executive role and ISO exercises absorb capacity), and run a partial-conversion calendar through the bridge years before RMD age 73. The alternatives considered: no conversion (rejected because RMD-age tax exposure on the $2.1M pre-tax will compound significantly), full conversion now (rejected because of 48.1% combined rate), partial conversion now sized to staircase below the 37% bracket (deferred because the household's facts suggest waiting yields a better outcome). IRC §408A and §401(a)(9) cited.

The Hendersons' Roth Conversion Memo

The 71-year-old retiree household post-inheritance. Same concept opening (the AI uses the same plain-English explanation). The memo's analysis specific to the Hendersons: combined household marginal rate is 24% federal + 3.8% NIIT + 0% state (Florida winter residency reduces effective state); pre-tax IRA is $1.6M post-inheritance accounting; the household is in RMD years (Mr. Henderson age 71 not yet, Mrs. Henderson age 73 is — RMD active); the IRMAA bracket capacity is the operative constraint; the household has documented charitable orientation post-inheritance (church expansion noted in May 2026 discovery). The memo's recommendation: a multi-year Roth conversion ladder sized to the IRMAA-capacity-floor under the 24% bracket fill, coordinated with QCD elections from the inherited IRA (IRC §408(d)(8) and §401(a)(9) cited) to reduce IRMAA-bracket exposure, plus a separate consideration of CRT funding for the charitable expansion. The alternatives considered: no conversion (rejected because the inherited IRA's 10-year-rule trajectory under SECURE 2.0 produces compressed tax exposure in years 4-10), full conversion (rejected because IRMAA bracket impact would dominate), the multi-year ladder (accepted as the household-best-interest path).

Why One Prompt Template Produces Two Different Memos

The prompt is identical. The inputs differ. The Wilsons' IPS, Holistiplan extraction, planning software, and meeting notes drive a fundamentally different recommendation than the Hendersons' inputs do. This is the lesson's central operational point: AI-generated personalization is not about generating "personalized-sounding" boilerplate. It is about taking household-specific structured inputs and producing household-specific reasoning, with citations, with alternatives, with a senior-advisor-verifiable recommendation. The prompt template is reusable; the output is irreproducible across households.

The "Clear and Prominent" Disclosure on AI-Generated Content

The January 2026 SEC staff FAQs did not relax the "clear and prominent" disclosure standard for material facts under Rule 206(4)-1(b). The firm's posture on AI-generated content varies in 2026; the practitioner consensus across Smarsh, Global Relay, Mintz, Sidley, and the L1 Ch4 L1 framework is that a "clear and prominent" disclosure of AI involvement should accompany any client-facing artifact materially drafted by AI when the firm's ADV Part 2A AI-use disclosure paragraph would not, by itself, put the household on notice that the specific artifact was AI-drafted.

A representative inline disclosure for an AI-generated quarterly commentary: "This commentary was drafted with the assistance of artificial intelligence tools and reviewed by the firm's senior advisor and compliance reviewer prior to distribution. Forward-looking statements reflect the firm's opinion at the date of publication and are not forecasts or guarantees. Past performance is not indicative of future results. This commentary is not a recommendation to buy or sell any security. Please refer to the firm's Form ADV Part 2A for material risks, fees, and conflicts."

A representative inline disclosure for an AI-generated concept memo: "This memo was drafted with the assistance of artificial intelligence tools using your specific household information (Investment Policy Statement, tax-return extraction, planning software output) and was reviewed by your senior advisor before delivery. The recommendations reflect your senior advisor's professional judgment. IRS code citations are accurate as of the date of publication; tax law is subject to change. This is not legal or tax advice — please consult your CPA or tax attorney before acting. The firm's Form ADV Part 2A discloses material risks, fees, and conflicts."

"Clear and prominent" is not a font-size rule; the SEC and Division will apply a holistic test. Inline, same-font, same-color placement adjacent to the content satisfies the standard; footer-only or 8-point-gray-text placement does not. The disclosure is the artifact the firm should standardize, version, and apply consistently across the book — not generated per-artifact by the AI.

The Hypothetical Performance Trap in Concept Memos

Concept memos describe future planning scenarios. The Wilsons' deferred Roth conversion memo, for example, may want to say "if you convert $200,000 per year for five years starting at age 63, you avoid approximately $X of RMD-era tax exposure based on assumptions A, B, C." This is hypothetical performance under Rule 206(4)-1(d). The rule generally prohibits hypothetical performance in general-audience advertisements unless audience-tailoring policies are in place. The household-specific concept memo is, arguably, "tailored to the audience" (the household itself) when delivered as a one-on-one written communication to a single advisory client — but the practical compliance posture in 2026 is to (a) frame projections as conditional examples rather than performance claims ("based on the stated assumptions, the projected RMD tax exposure is approximately $X"), (b) explicitly list the assumptions and their sources, (c) note the limitations of the projection (tax law changes, market changes, household-facts changes), and (d) avoid presenting the projection as a guarantee of outcome. The L4 Ch7 lesson develops the hypothetical-performance policy framework; this lesson applies the conditional-example framing.

The Pre-Use Review Queue and the Archive

Every AI-drafted quarterly commentary and every AI-drafted concept memo passes through the firm's pre-use review queue under FINRA Rule 2210 (for hybrid practices) and Marketing Rule substantiation review (for RIAs). The L4 Ch3 architecture handles the queue at scale: risk-based sampling, AI-to-AI red-team review for first-pass screening, and exception handling for the artifacts that fail substantiation, disclosure, or balance tests. The senior advisor's review pass is the human checkpoint that catches the residual issues the AI's first-pass screening misses.

The archive bundle: prompt, AI tool / model version, AI-generated draft, advisor edits, reviewer signoff, distribution date, recipient list (for the commentary) or recipient identity (for the concept memo). Retained per SEC Rule 204-2 (5 years, first 2 in easily accessible location) and FINRA Rule 4511 (3 years for BD records). The Smarsh or Global Relay archive holds the bundle; the L3 Ch10 L1 Zocks-to-Wealthbox-to-Smarsh pipeline lesson handles the meeting-AI-side archiving, and the L4 Ch3 lesson handles the principal-review-queue side.

The 90-Second Client Framing

The household-facing explanation, refined across hundreds of advisor conversations: "You'll receive two kinds of written content from us. The first is a quarterly market commentary that's substantially the same for every household — our view of what happened in the market and what we're watching. The second is a concept memo when something specific to your situation comes up — a Roth conversion window, a QCD opportunity, an NUA decision, a 72(t) calculation. The quarterly commentary is drafted with AI assistance and reviewed by our team before distribution; the concept memos are personalized to your specific household data and are also reviewed before they reach you. Every word that goes to you is a regulated communication under the SEC Marketing Rule and our firm's compliance program. If you have questions about how we use AI in producing these, our Form ADV Part 2A explains it; if you want any of it sent in a different format or not at all, just let us know."

Key Takeaways

• Quarterly commentary and personalized concept memos are both advertisements under SEC Rule 206(4)-1; the 2024-2025 AI-washing settlements, January 2026 staff FAQs, and FINRA 2026 Oversight Report frame the operational expectations.
• Quarterly commentary uses a five-section structure (market summary, macro context, forward-looking opinion, IC tilts within IPS ranges, standing disclosures) — forward-looking-statement-safe framing, no cherry-picked performance, every macro data point cited.
• Personalized concept memos use one prompt template across all households — the household-specific inputs (IPS, Holistiplan extraction, planning software, meeting notes) drive fundamentally different recommendations for a 58-year-old executive vs. a 71-year-old retiree on the same concept.
• "Clear and prominent" disclosure for AI-generated content — inline, same-font, same-color, adjacent to the content — is the firm-standardized artifact, not a per-output AI generation. The January 2026 staff FAQs did not relax this standard.
• Hypothetical performance under Rule 206(4)-1(d) in concept memos is handled via conditional-example framing with stated assumptions, sources, and limitations — not as a guaranteed projection.
• Every output passes through the L4 Ch3 pre-use review queue (FINRA Rule 2210 for hybrid practices, Marketing Rule substantiation for RIAs) and the L4 Ch7 substantiation file. The archive (Smarsh / Global Relay) holds the prompt, model version, draft, edits, signoff, date, and recipient detail for the longer of Rule 204-2 / Rule 4511 periods.
• Cited IRC sections in concept memos: §408A Roth IRA rules, §401(a)(9) RMD coordination, §72(t) early-withdrawal penalty, §402(e)(4) NUA, §408(d)(8) QCDs, §409A deferred comp — each pulled into the memo when the household's facts call for it.