AI financial advice: when your app becomes an unregistered investment adviser

The line between a 'financial information tool' and an 'unregistered investment adviser' is narrow and the SEC has been actively enforcing it against AI-flavoured retail products. The Investment Advisers Act's three-part test (advice, securities, compensation) catches a lot of products founders didn't think were in scope. Here's the test, the Marketing Rule, the enforcement record, and the registration-vs.-disclaimer decision every operator has to make.

The three-part test

Section 202(a)(11) of the Investment Advisers Act of 1940 defines an 'investment adviser' as any person who, for compensation, engages in the business of advising others as to the value of securities or as to the advisability of investing in, purchasing, or selling securities. The SEC's published interpretation (Release IA-1092, 1987) reduced this to a practical three-part test.

1. Advice — does the product provide recommendations about securities or financial planning?
2. Securities — are the recommendations about stocks, bonds, mutual funds, ETFs, crypto-securities, options, or similar instruments?
3. Compensation — does the operator receive compensation (subscription fees count; ad revenue counts; affiliate commissions count)?

If your product answers "yes" to all three, the default legal status is "investment adviser, registration required." The exceptions are narrow: publishers (the Lowe v. SEC line — bona fide newsletters with general circulation), pure educational content, and certain de minimis exemptions. Most AI 'advisor' or 'robo-advisor' products don't fit the exceptions.

The SEC Marketing Rule

Even before registration questions, the SEC Marketing Rule (Rule 206(4)-1, updated 2022, fully in effect 2024) governs how registered AND unregistered advisers can talk about their service. Several recent SEC actions cite Marketing Rule violations by AI-flavoured products as the operative violation.

• No guaranteed returns. "Risk-free", "guaranteed", "certain profit" are per-se misleading under the rule.
• No cherry-picked performance. If you show backtested or hypothetical returns, you must show the methodology and limitations prominently. "Beat the market with AI" claims without GIPS-compliant backing are the typical Marketing Rule trigger.
• No testimonials without disclosure of compensation and material conflicts.
• AI-specific scrutiny — the SEC's 2024 'AI-washing' enforcement sweep targeted multiple firms claiming AI capabilities they didn't have. Don't overstate the AI component.

The enforcement record against AI advice products

• SEC v. Delphia (2024) — registered RIA, $225K Marketing Rule settlement for AI-capability overclaim
• SEC v. Global Predictions (2024) — registered RIA, $175K Marketing Rule settlement for similar overclaim
• FINRA actions against retail crypto-recommendation tools (multiple, 2024-2025) — typically for failure to register, performance claim violations, or both
• State securities regulators — NASAA (the state-regulator association) issued a 2024 alert specifically targeting 'AI investment advice' products, which presages coordinated state enforcement

Language that triggers SEC / FINRA scrutiny

The patterns below appear in nearly every enforcement action against software-mediated investment products. If your product copy includes these phrases, you're operating well above baseline regulatory risk.

• "Based on your goals/situation/portfolio, you should…" — personalisation is the operative test for being an adviser
• "Guaranteed returns" / "risk-free" / "can't lose" — per-se Marketing Rule violations
• "Beat the market" / "alpha" / "outperform the S&P" without GIPS-compliant backing — Marketing Rule
• "AI-powered investing" / "proprietary AI model" without disclosed methodology — AI-washing risk
• "Insider tips" / "sure thing picks" — appears in nearly every SEC fraud action against retail products

The registration vs. disclaimer fork

Every AI-financial product eventually faces the same decision: register as an investment adviser, or restructure to avoid registration. There's no third path that scales.

Path A — register

SEC registration if AUM > $100M, state registration (one or many) below that threshold. Form ADV filing, fiduciary duty obligations, recordkeeping (Books and Records Rule), compliance officer requirement, custody rule, advertising rule, code of ethics, business continuity plan. Initial filing + first-year compliance setup typically $25K–$75K for an SMB; ongoing compliance $15K–$50K/year. Once registered, you can give personalised advice freely.

Path B — restructure to avoid registration

Reframe outputs as educational analysis rather than recommendations. Add 'not investment advice' disclaimers. Avoid personalisation language. Use 'considerations' and 'characteristics to look at' instead of 'buy / sell'. Add explicit 'we are not a registered investment adviser' disclosure. This is the Morningstar / Seeking Alpha / Motley Fool model.

The educational-only floor

If you go the disclaimer route (Path B), the floor that has worked for the established players is:

1. Persistent disclaimer in footer: "This service is for informational and educational purposes only. It is not investment advice, not a recommendation to buy or sell any security, and not personalised to your situation."
2. Per-output disclaimer: "Educational analysis. Not investment advice." Above every model output that mentions a security.
3. Registration-status disclosure: "We are not a registered investment adviser or broker-dealer."
4. No personalisation language. Replace "you should" with "products with these characteristics include".
5. Past-performance disclaimer wherever historical returns appear: "Past performance does not predict future results."

What to do this week

1. Apply the three-part test honestly. If you answer 'yes' to all three, decide registration vs. restructure now — the cost of doing it later is materially higher.
2. Audit every output for personalisation language ('based on your', 'you should', 'we recommend'). Replace with educational framing OR commit to registering.
3. Audit every marketing claim for AI-washing exposure. Don't claim AI capabilities you can't substantiate.
4. Add the registration-status statement to your footer. Either claim registration (with Form ADV link) or explicitly disclaim it.
5. Re-scan after each change. Comply Code's financial rule pack flags personalisation language, guaranteed-return phrases, missing disclaimers, and missing registration-status disclosure.