TL;DR: SEC AI Governance for Investment Advisers: 2026 Examination Priorities, a practical compliance guide for enterprise and HR teams in 2026.
The SEC's Division of Examinations published its 2026 examination priorities in January 2026, and AI oversight appeared in the document for the second year in a row. For registered investment advisers (RIAs) and broker-dealers, this means that when examination staff walk through your door, they will ask specific questions about how you are using AI, whether your disclosures are current, and whether your AI systems create conflicts of interest with your clients.
This is no longer theoretical. Examination staff received training on AI-related examination procedures throughout 2025, and the 2026 priorities document makes explicit that examiners will focus on AI in compliance monitoring, client communications, investment research, and trading. If you are using AI and have not structured your governance program to respond to these examinations, you have a gap that needs closing before your next exam cycle.
This guide covers what examiners are asking, which uses of AI are in scope, the conflict-of-interest risk the SEC has identified, Form ADV disclosure requirements, and a 10-point examination readiness checklist.
What the 2026 examination priorities say about AI
The Division of Examinations' annual priorities letter does not merely mention AI in passing. The 2026 document identifies AI as a priority area for examinations of investment advisers and broker-dealers, with examination staff directed to assess whether firms have appropriate policies, procedures, and controls governing AI use.
The specific areas the priorities document calls out include: AI used in investment advice or investment decision-making processes; AI used in compliance programs; AI used in client-facing communications or marketing materials; and AI used in trading or order management. The document also flags the integrity of data inputs to AI systems as a concern, meaning examiners may ask about data quality and provenance, not just about the AI models themselves.
The 2026 priorities also continue the Division's focus on conflicts of interest arising from technology. This was the central concern in the September 2023 proposed rule on predictive data analytics and conflicts of interest (Release No. IA-6383), which would have required advisers using covered technology to eliminate or neutralize conflicts of interest favoring the adviser over clients. While the rule has not been finalized as of mid-2026, the underlying concern drives examination questions under existing fiduciary duty requirements.
Examiners are also looking at disclosures. Firms that began using AI tools after their last Form ADV update and have not amended their disclosures are at particular risk of findings related to inaccurate or incomplete disclosure.
Which AI uses are in scope
Not every use of software counts as "AI" for examination purposes, but the SEC has cast a wide net in its communications. The proposed predictive data analytics rule defined covered technology as "analytical, technological, or computational functions, including, but not limited to, algorithms or models that optimize for, predict, guide, forecast, or direct investment-related behaviors or outcomes." That definition covers most modern recommendation systems, analytical tools, and automation with any optimization component.
Robo-advisory platforms are the most obvious in-scope use. Any system that generates investment recommendations using algorithms or machine learning falls squarely within the SEC's focus. This includes third-party robo-advisory infrastructure that an RIA licenses, not just proprietary systems. If you are a human adviser who uses a robo-advisory backend to generate model portfolios, that counts.
AI-generated investment research is in scope. If you use a tool like a large language model to summarize earnings reports, generate investment theses, or screen securities, that is an AI use in your investment process that examiners may inquire about. The question is not just whether the tool has appropriate disclosures, but whether the research it generates creates any conflicts of interest.
AI in compliance monitoring is in scope and potentially ironic given that it is used to support compliance. If you use AI to monitor communications for regulatory violations, screen employee trades, or flag suspicious activity reports, examiners may ask about the configuration of those systems and whether they are calibrated to catch problems or to miss them.
AI in client communication includes AI-generated newsletters, AI-drafted performance reports, AI chatbots on your client portal, and AI tools used to respond to client inquiries. If any client-facing communication is touched by AI, that should be in your AI use inventory.
AI-assisted trading covers any system that automates order routing, timing, or sizing using algorithmic or AI-based logic, even where a human approves the final order.
The conflict-of-interest risk
The SEC's core concern with AI in investment advice is that optimization systems can systematically favor adviser interests over client interests in ways that are not transparent and may not even be intentional. This is the scenario the 2023 proposed rule was designed to address, and it remains the animating concern behind examination questions even without a final rule.
The clearest conflict-of-interest scenario involves recommendation systems. An AI system trained or fine-tuned on historical data where the firm's more profitable products are recommended more often will tend to recommend those products more often going forward. The optimization objective may have been benign (recommend what worked in the past), but the output may systematically tilt toward products that earn the firm higher revenue. Under the fiduciary duty codified in the Investment Advisers Act of 1940, this is a problem regardless of whether anyone intended it.
A second conflict scenario involves data inputs. If the training data or reference data for an AI research tool is skewed, for instance, if it over-weights research from analysts who cover companies that are also investment banking clients, the outputs may systematically favor those companies. The fiduciary duty applies to the outputs the adviser relies on, not just to intentional recommendations.
A third scenario involves AI in compliance monitoring. A compliance monitoring AI configured with threshold settings that are calibrated conservatively (avoiding false positives) may miss actual violations. If that calibration was influenced by a desire to avoid compliance findings, that could be characterized as a conflict of interest in how compliance resources are deployed.
To address these risks, the SEC expects advisers to be able to demonstrate that they have evaluated their AI systems for potential conflicts of interest, documented those evaluations, and either eliminated the conflict or disclosed it in Form ADV. This requires that someone at your firm understand the optimization objective of each AI system you use and can articulate what that optimization is maximizing.
Form ADV disclosure requirements
Form ADV Part 2A, the client brochure, is the primary disclosure vehicle for investment advisers. Item 8 covers methods of analysis, investment strategies, and risk of loss. Item 10 covers financial industry affiliations. Item 12 covers brokerage practices. Depending on how you use AI, disclosures may be required in one or more of these items.
For AI used in generating investment recommendations or research, the disclosure in Item 8 should describe the role of AI in your investment process, the types of AI tools or systems used (at a level of specificity that is meaningful without disclosing proprietary information), and any material limitations of those tools. If the AI system uses specific data sources that could create conflicts of interest, those should be disclosed.
For AI used in compliance, no mandatory Form ADV disclosure item specifically covers compliance technology, but if the compliance AI is relevant to how you carry out your supervisory responsibilities, it may be appropriate to reference it in Item 9 (disciplinary information) or in narrative disclosures about your compliance program.
For AI-generated client communications, disclosure in the client brochure should note that client communications may be prepared with AI assistance, and that human review occurs before distribution. If communications are fully automated without human review, that should be disclosed clearly.
The annual amendment deadline for Form ADV is 90 days after your fiscal year end. However, if material changes occur mid-year, you must promptly update Form ADV under Rule 204-1 and notify clients of material changes. If you launched an AI-based investment tool in Q1 and your last Form ADV predates that launch, your disclosure is materially incomplete and should be updated without waiting for the annual amendment cycle.
10-point examination readiness checklist
Use this checklist to assess your current posture before your next examination.
-
Complete an AI inventory. List every AI tool in use at your firm, who uses it, what it is used for, and whether it touches investment decisions, compliance, or client communications. Your AI governance checklist 2026 is a starting point for the framework, but investment advisers need the additional coverage addressed here.
-
Review Form ADV for completeness. Compare your current Form ADV disclosures against your AI inventory. For every AI use not currently disclosed, determine whether it is material and draft language for the next amendment if it is.
-
Document the optimization objective of each AI system. For each AI tool in your investment process or compliance program, document in writing what the system is optimizing for, what data it uses, and who configured it. This is the documentation examiners will ask for.
-
Evaluate each AI system for potential conflicts of interest. For each tool, assess whether its optimization objective could favor the firm over clients. If a conflict exists, determine whether you will eliminate it, mitigate it, or disclose it. Document the evaluation and conclusion.
-
Check vendor contracts and disclosures. For third-party AI tools, review the vendor's terms of service and data handling practices. Do they use your clients' data to train their models? If so, that may need to be disclosed. See the AI vendor due diligence checklist 2026 for the specific questions to ask.
-
Review your policies and procedures for AI. Investment advisers registered with the SEC must have written compliance policies and procedures under Rule 206(4)-7. Those policies must be reviewed annually. If your policies do not address AI use, add an AI governance section before your next annual review.
-
Verify data quality and provenance for AI systems used in investment decisions. Identify the data sources feeding each AI system. Assess whether those sources could introduce bias or conflicts of interest. Document your assessment.
-
Test AI-generated client communications. Pull a sample of recent client communications that were drafted or modified by AI. Review them for accuracy, consistency with your investment approach, and compliance with applicable rules including the Marketing Rule (Rule 206(4)-1).
-
Check for AI use in advertisements and solicitations. The SEC Marketing Rule requires that performance figures and testimonials in adviser marketing meet specific requirements. AI-generated marketing copy must be reviewed to confirm it does not include prohibited content such as untrue statements of material fact or misleading performance presentations.
-
Brief your principals and examination team. When examiners arrive, they may ask your principals directly about AI use. Your CEO, CIO, CCO, and any portfolio manager using AI tools should be able to describe those tools and your governance framework. Examiners are trained to identify inconsistencies between what the written policies say and what staff say during interviews.
Preparing for examiner questions
The SEC's Division of Examinations typically begins a routine examination with a document request letter, followed by on-site or remote meetings. For AI-related topics, expect document requests to include your AI tool inventory or register, any written policies governing AI use, Form ADV disclosures, vendor contracts for AI tools, and any internal evaluations of AI systems for conflicts of interest.
During interviews, examiners may ask your CCO to describe how the firm monitors AI systems for compliance risks. They may ask portfolio managers to describe what AI tools they use and how they factor AI outputs into investment decisions. They may ask your IT or operations team about how AI outputs are logged and retained.
The key to a clean examination is consistency. What your policies say, what your Form ADV discloses, what your staff says in interviews, and what your systems actually do should all align. Discrepancies between policy and practice are a common finding in SEC examinations, and AI is now a topic where those discrepancies are being actively probed.
The SEC's 2026 examination priorities publication is available on the SEC website. The 2023 predictive data analytics proposed rule (Release No. IA-6383) provides the detailed conceptual framework for how the SEC thinks about AI conflicts of interest, even though the rule has not been finalized. Both documents are worth reading if you want to understand the regulatory mindset your next examination team will bring.
Building an AI governance program that satisfies SEC examination standards is not a one-time project. It requires ongoing documentation, annual policy review, and regular assessment of new AI tools as they are adopted. The AI regulation deadline calendar 2026 can help you track the SEC examination cycle alongside other regulatory timelines that apply to your firm.
