L5 Ch3 L1 produced the enterprise AI policy. L5 Ch3 L2 produced the data architecture and consent management framework. This lesson produces the client-facing posture — what the firm tells the client, what the client gets to choose, and what the firm defaults. The client-facing posture in 2026 is not a single conversation; it is a layered disclosure structure spanning the engagement letter, ADV Part 2A, Form CRS, the client portal, the meeting introduction, and the trade authorization flow — each anchored to the SEC Marketing Rule 206(4)-1 "clear and prominent" standard, the Reg BI §240.15l-1 disclosure obligation, FINRA Rule 2210, NAIC Model #275 for the annuity-licensed channel, and the state DOI patchwork per L4 Ch6 L2. This lesson installs the framework and the 2027-2028 trajectory the firm should be planning against now — based on SEC, state, and NAIC rulemaking signals already in motion.

The Three Disclosure Categories — What's Told, Opt-Ed, Defaulted

Told — Mandatory Disclosure

Certain AI uses must be disclosed regardless of client preference. The mandatory disclosure category includes: the firm uses AI tools in advisor workflow (engagement letter + ADV Part 2A AI disclosure paragraph per L5 Ch3 L1); the firm's recommendations are made by registered individuals exercising fiduciary judgment under the Investment Advisers Act of 1940 and Reg BI §240.15l-1, not by AI (the L1 Ch1 framing); the firm's Reg S-P 17 CFR Part 248 client-information handling extends to AI vendors; client data may be processed by approved vendors under executed data-handling agreements; ADV Part 2A annual updating amendment + off-cycle prompt amendments under IA-1992 cover material AI tool changes affecting data handling; Marketing Rule 206(4)-1 substantiation file per L4 Ch7 L1 supports any external AI claim; Form U4 disclosure obligations under FINRA Rule 4530 apply to registered persons.

Opt-In By Default — Client Affirmative Choice

Certain AI capabilities require affirmative client opt-in. The opt-in category includes: agentic AI action-taking per L4 Ch3 L3 framework (trade execution, rebalancing to IPS, beneficiary maintenance, RMD processing, ACATs initiation, draft 1040 amendments — anything where AI takes action on the client's behalf); client portal AI concierge features per L5 Ch5 L1 (24/7 client-facing AI that can answer questions and route to advisor); AI use of client-specific health information for planning (longevity assumptions, long-term-care projections); AI cross-household pattern mining if any client identifying information is involved. Opt-in is captured in the engagement letter as a specific affirmative check; revocable at any time; documented in the CRM custom field; propagated through the integration layer per L5 Ch3 L2.

Defaulted — Firm Operating Model

Certain AI uses are part of the firm's operating model and apply to all clients unless the client requests specific exclusion. The defaulted category includes: meeting capture and summarization (Jump, Zocks); document extraction (Holistiplan for tax returns, FP Alpha for estate documents, Wealth.com for estate); draft generation of client communications subject to principal review under Rule 2210; pattern detection in transaction histories; CRM next-best-action prompts; internal research support; prompt-library iteration. Default is governed by the engagement letter's broad consent language and the ADV Part 2A AI disclosure. Client may request specific tool exclusion (e.g., "do not use meeting AI in my reviews") captured in the CRM custom field with documented Reg BI Care Obligation rationale per L5 Ch3 L2.

The Layered Disclosure Structure

Engagement Letter — The Foundation

The engagement letter establishes the broad consent for the firm's AI operating model and the affirmative opt-in mechanisms for agentic AI and portal AI. The 2026 standard structure includes: a Section X on AI use describing the firm's operating model in plain language; a checkbox or signature line for engagement-letter consent to supporting AI; separate checkboxes or signature lines for agentic AI opt-in (with per-action-category granularity if appropriate); and a reference to the ADV Part 2A and Privacy Policy for further detail. Engagement letter updates are required on material AI tool changes per L5 Ch7 L6 ADV-diff workflow if the changes affect engagement scope.

ADV Part 2A — The Regulatory Disclosure

The ADV Part 2A AI disclosure paragraph per L5 Ch3 L1 names the AI tool categories (meeting AI, tax-extraction, estate-extraction, CRM AI, planning AI), anchors recommendations to registered individuals under the Investment Advisers Act of 1940 fiduciary duty and Reg BI §240.15l-1, references Rule 2210 principal review of AI-drafted content, and references Reg S-P 17 CFR Part 248 client-information handling. Annual updating amendment per L5 Ch7 L6; off-cycle prompt amendments under IA-1992 for material changes mid-year. The substantiation file per L4 Ch7 L1 backs any external claim about AI capability.

Form CRS — The Plain-Language Disclosure

Form CRS (Client Relationship Summary) Item 1 (Relationships and Services) and Item 5 (Disciplinary History) carry the most plain-language AI references. Item 1 mentions AI use in the firm's services description if material; the language is consistent with the ADV Part 2A paragraph but more compressed; the "Conversation Starters" Item 4 may include AI-specific questions. Material AI changes affecting Form CRS trigger redelivery to existing clients per SEC requirements.

Client Portal Disclosure

The L5 Ch5 L1 client portal carries layered disclosure: a banner or footer indicating AI involvement in any AI-generated content; specific "AI-assisted" labels on AI-drafted summaries, recommendations, or analysis; opt-in flow for agentic features (with affirmative consent capture); audit trail of client interactions for FINRA Rule 4511 retention. The portal's AI is disclosed at first use and at any material upgrade.

Meeting Introduction — The Verbal Disclosure

The advisor's verbal meeting introduction acknowledges meeting AI use at first encounter and at annual reviews. The 2026 standard 30-second script: "I'm using a meeting AI tool to capture today's conversation — it helps me focus on you rather than note-taking. I review the output before anything goes in your file or any communication goes out. If you'd prefer I not use the tool for any reason, let me know now or anytime. Our engagement letter and Form ADV describe the firm's full AI use; happy to discuss any of it."

Trade Authorization — The Decision-Point Disclosure

When AI is involved in producing a trade authorization (e.g., a rebalance memo drafted by AI), the trade authorization flow discloses the AI involvement at the decision point: "This recommendation was drafted with AI assistance and reviewed by your advisor. Click to review the recommendation and approve, modify, or decline." The disclosure is at the moment of client consent for the action, complementing the broader engagement-letter and ADV-level disclosures.

Agentic AI Disclosure Architecture

Agentic AI requires the most explicit disclosure architecture because the AI takes action rather than producing drafts. The 2026 baseline per L4 Ch3 L3 + L5 Ch3 L1 is opt-in for each agentic capability category.

Trade Execution Agentic AI

AI initiating trades within IPS bounds and pre-approved parameters. Disclosure: opt-in checkbox in engagement letter with per-account-type granularity (e.g., IRA only, taxable only, both, none); kill-switch architecture per L4 Ch3 L3; post-action review queue staffed under Rule 2210; client receives confirmation of each agentic trade with notation that AI initiated; client may request review and human signoff for any specific trade or trade category at any time.

Rebalancing Agentic AI

AI rebalancing portfolios to IPS targets within tolerance bands. Disclosure: opt-in with IPS reference; client receives quarterly rebalancing summary including AI-initiated transactions; principal-review queue under Rule 2210 covers any deviation from IPS bounds; Reg BI Care Obligation §240.15l-1 documentation for the rebalancing recommendation framework remains the registered individual's responsibility.

RMD Processing and ACATs Initiation

AI processing RMDs by deadline and initiating ACATs transfers. Disclosure: opt-in for RMD processing with annual review of calculations; opt-in for ACATs with explicit per-transfer authorization. The IRC 401(a)(9) RMD calculation accuracy is verified per Cardinal Rule L1 Ch2.3 source-system tier before action; the SECURE 2.0 73-vs-75 confusion is the classic hallucination risk to guard against.

Beneficiary Maintenance

AI maintaining beneficiary designations across accounts. Disclosure: opt-in with explicit warning that beneficiary updates are legally significant and benefit from human review; client confirms each update before AI processes; estate-planning interaction handled per L3 Ch5 L2 estate gap and beneficiary audit workflow.

The 2027-2028 Disclosure Trajectory

Based on current SEC, state, and NAIC rulemaking signals, the 2027-2028 disclosure trajectory likely includes several specific developments. The lesson plans the firm's roadmap against these signals rather than waiting for final rules.

SEC Trajectory

The SEC's evolving cybersecurity rulemaking is likely to formalize cybersecurity incident disclosure requirements for advisers, building on the May 2024 Reg S-P 17 CFR Part 248 amendments and the FINRA 2026 Annual Regulatory Oversight Report's framing. The Marketing Rule 206(4)-1 staff FAQs are likely to expand on AI-specific claims, third-party-rating mechanics, and hypothetical-performance under Rule 206(4)-1(d) — building on the January 2026 FAQs and the 2024-2025 Delphia/Global Predictions enforcement. The SEC may also clarify the relationship between AI use and ADV Part 2A material-change disclosure under IA-1992, formalizing when AI tool changes trigger off-cycle prompt amendments.

State DOI Trajectory

The NAIC AI Model Bulletin (March 2026 issue brief) and Model #275 (annuity suitability and best-interest) are likely to produce more state-by-state adoption with AI-specific disclosure requirements layered on. The L4 Ch6 L2 50-state matrix gains complexity as states adopt variations. NY DFS 23 NYCRR 500 is likely to issue AI-specific guidance under the third-party-service-provider framework, formalizing expectations on agentic AI cybersecurity. State CPRA / DIR / equivalents will continue adding AI-specific provisions.

FINRA Trajectory

The FINRA 2026 Annual Regulatory Oversight Report framed agentic AI as a Rule 3110 reasonable-design issue and prompts-as-records under Rule 4511. The 2027 and 2028 reports are likely to refine these frames with more specific WSP expectations, additional Rule 2210 principal-review guidance for AI-drafted content, and clearer Rule 4530 customer-complaint handling for AI-involved matters. Form U4 narrative drafting per L5 Ch7 L4 may evolve as FINRA provides more agentic-AI-specific examples.

Client Experience Trajectory

The 2026 baseline of "engagement letter + ADV + meeting introduction + portal banner" disclosure structure is likely to evolve toward more specific per-interaction disclosures by 2027-2028. Clients may expect to see AI involvement labeled in every output (every recommendation, every analysis, every portal display) rather than just at first contact. The firm's L5 Ch5 L1 client portal architecture should plan for this evolution. The L5 Ch5 L2 pricing and service-tier framing may shift if AI-augmented service tiers become distinctly marketed.

Case Study — Network-Wide Disclosure Framework Rollout

A $7B RIA network with 130 advisors rolled out its layered AI disclosure framework in Q4 2025-Q1 2026. The framework: engagement letter Section X on AI use with three checkbox tiers (supporting AI consent, agentic AI opt-in by category, portal AI opt-in); ADV Part 2A AI disclosure paragraph aligned to L5 Ch3 L1 standard; Form CRS Item 1 updated for AI; client portal banner and per-output labels; meeting introduction 30-second script trained across 130 advisors; trade authorization flow with AI-assistance notation. The rollout: 60 days drafting + outside counsel review; 60 days advisor training + practice meeting introductions; 30 days client communication and engagement letter refresh for active clients; 30 days portal deployment; total 6 months.

The Q2 2026 outcomes: 92% of active clients refreshed engagement letter; 22% of clients opted in to portal AI; 8% opted in to trade-execution agentic AI; 4% opted in to rebalancing agentic; 2% opted in to RMD agentic; the broader pattern shows clients comfortable with supporting AI (98% no specific exclusions) but cautious on agentic. Reg S-P 17 CFR Part 248 incidents: zero AI-disclosure-related. Marketing Rule 206(4)-1 audit per L4 Ch7 L1: clean. L4 Ch8 L1 supervisory-architecture score: 8/10 → 9/10. The firm's outside counsel framed the disclosure framework as defensible across SEC, FINRA, state DOI, and NAIC AI Model Bulletin expectations.

Key Takeaways

• Client-facing AI disclosure has three categories: told (mandatory under Reg BI §240.15l-1, Marketing Rule 206(4)-1, ADV Part 2A, Reg S-P 17 CFR Part 248, Form U4); opt-in (agentic action-taking per L4 Ch3 L3, portal AI concierge per L5 Ch5 L1, health data, cross-household pattern mining); defaulted (firm operating model — meeting capture, document extraction, draft generation, pattern detection — with client-requested tool exclusion).
• The layered disclosure structure spans engagement letter (foundation with opt-in checkboxes), ADV Part 2A (regulatory disclosure annually updated per L5 Ch7 L6 with off-cycle prompt amendments under IA-1992), Form CRS (plain-language Item 1 and Item 4 conversation starters), client portal (banner + per-output labels), meeting introduction (verbal 30-second script), trade authorization (decision-point disclosure).
• Agentic AI disclosure is opt-in by category: trade execution with per-account-type granularity, rebalancing with IPS reference, RMD processing with annual calculation review (SECURE 2.0 73-vs-75 hallucination guard via Cardinal Rule L1 Ch2.3 source-system tier), ACATs initiation with per-transfer authorization, beneficiary maintenance with explicit legal-significance warning. Each opts-in independently; kill-switch architecture and post-action review per L4 Ch3 L3 supervise.
• The 2027-2028 trajectory: SEC likely formalizes cybersecurity incident disclosure under evolving rulemaking and expands Marketing Rule 206(4)-1 AI-specific FAQs; state DOI continues adopting NAIC AI Model Bulletin variations with L4 Ch6 L2 50-state matrix gaining complexity; FINRA 2027/2028 reports refine Rule 3110 agentic-AI WSP expectations and Rule 2210 principal-review for AI-drafted content; client experience evolves toward per-interaction disclosure labels.
• Case study: $7B network with 130 advisors rolled out 6-month framework. 92% engagement letter refresh; 22% portal AI opt-in; 8% trade-execution agentic opt-in. Zero Reg S-P AI-disclosure incidents; clean Marketing Rule audit; L4 Ch8 L1 supervisory-architecture score 8/10 → 9/10. Outside counsel confirmed defensibility across SEC, FINRA, state DOI, and NAIC AI Model Bulletin.
• The firm's posture is durable when each disclosure layer is consistent with the others. Engagement letter language matches ADV paragraph matches Form CRS matches portal banner matches meeting script matches trade-authorization flow. Inconsistency between layers creates Marketing Rule 206(4)-1 substantiation failure and ADV currency exposure. The L4 Ch7 L1 audit + L5 Ch7 L6 ADV-diff workflow + L5 Ch3 L1 enterprise policy together maintain consistency.