The Advisor's Guide to Using AI Without Compliance Headaches

Every advisor we talk to has the same concern about AI: "My compliance department will never approve this."

We get it. Financial services is one of the most heavily regulated industries, and the idea of using AI to help with client communications feels like walking into a minefield.

But here's the reality: AI isn't going away. Your competitors are already using it. The question isn't whether to use AI—it's how to use it responsibly in a regulated environment.

After months of working with AI tools in our own RIA practice (and plenty of conversations with compliance consultants), here's the practical framework we've developed.

The Core Principle: AI Assists, Humans Decide

The single most important thing to understand is that AI should never make decisions for clients or generate final client-facing content without human review.

Think of AI as a very capable intern. You wouldn't let an intern send client emails without your approval. You wouldn't let them make investment recommendations. You wouldn't let them sign off on financial plans.

The same rules apply to AI. It can draft, research, summarize, and suggest—but a licensed professional must review and approve everything before it reaches a client.

What's Safe vs. What's Risky

Generally Safe Uses

These are activities where AI can add significant value with relatively low compliance risk:

• Drafting client emails — AI writes a first draft, you review and personalize before sending
• Summarizing meeting notes — AI extracts key points from transcripts for your CRM
• Research assistance — AI helps you understand investment products, tax strategies, or market conditions
• Internal documentation — Meeting prep, internal memos, process documentation
• Content creation — Blog posts, newsletter drafts, educational materials (with review)
• Explaining concepts — Having AI help you craft explanations of complex topics

Higher Risk Uses (Proceed with Caution)

These activities require more careful consideration and potentially compliance pre-approval:

• Investment recommendations — AI should never generate specific buy/sell recommendations
• Financial projections — Any numbers need verification; AI can make calculation errors
• Regulatory filings — ADV updates, Form CRS modifications, etc.
• Personalized advice — Anything that could be construed as specific financial advice to a specific client

Data Privacy: What Can You Share with AI?

This is where many advisors get stuck. The answer depends on which AI tool you're using and how it handles data.

Claude (Anthropic)

Claude Pro does not use your conversations to train its models. Your conversations are retained for 30 days for trust and safety purposes, then deleted. Anthropic also offers a business tier with additional privacy controls.

ChatGPT (OpenAI)

By default, OpenAI may use your conversations for model training. You can opt out in settings, but many compliance departments are more comfortable with tools that don't train on user data by default.

Best Practices for Any AI Tool

A practical approach: Share the situation, not the identity. Instead of "John Smith has $2.3M in his IRA and is worried about RMDs," try "A 72-year-old client with a significant IRA balance is asking about RMD strategies. Help me explain the options."

Documentation: Covering Your Bases

If your compliance department asks "Did you use AI to write this?", you should have a clear answer and documentation.

What to Document

• The type of task — Was this a draft email, meeting summary, research, etc.?
• Human review occurred — Note that you reviewed and approved the output
• Any modifications made — Document significant changes from the AI draft

Simple Documentation Approach

We keep it simple: If AI helped with something client-facing, we add a note to the CRM record. Something like: "Email drafted with AI assistance, reviewed and personalized by [advisor name]."

This isn't currently required by regulation, but it's good practice and shows good faith if questions ever arise.

Building a Firm Policy

If you're a solo advisor, your "policy" can be your own guidelines. If you have a team or work under a broker-dealer, you'll want something more formal.

What Regulators Are Saying

As of early 2026, there's no specific SEC or FINRA rule about advisors using AI tools. But that doesn't mean it's the Wild West.

Existing regulations still apply:

• Fiduciary duty — You're still responsible for advice, regardless of how it was generated
• Books and records — Client communications must still be retained
• Advertising rules — AI-generated marketing content must comply with existing requirements
• Supervision — Firms must supervise AI use like any other tool

The regulatory landscape is evolving. The SEC has signaled interest in AI-related risks, and we expect more specific guidance in the coming years. Building good practices now positions you well for whatever comes.

Practical Tips from Our Experience

Start Small

Don't try to automate everything at once. Start with internal tasks like meeting summaries or research. Build confidence before moving to client-facing communications.

Always Review Output

Even for simple tasks, read what the AI produces. Look for factual errors, tone problems, or anything that doesn't sound like you.

Keep Compliance in the Loop

If you work under a broker-dealer or have an external CCO, have a conversation about AI use before diving in. It's easier to get buy-in upfront than to ask for forgiveness later.

Train Your Voice

The better you train AI to write in your voice, the less editing you'll need—and the less likely clients will notice AI involvement. Check out our guide on training Claude to write like you.

Stay Current

AI capabilities and regulatory guidance are both evolving rapidly. What's true today may change in six months. Follow industry publications and compliance bulletins.

The Bottom Line

Using AI in your advisory practice isn't inherently risky—using it carelessly is. With proper guardrails, human oversight, and documentation, AI can make you more efficient without creating compliance problems.

The advisors who figure this out now will have a significant advantage over those who wait until it's "perfectly clear" what the rules are. That clarity may take years.

Start with low-risk use cases, document your process, and always remember: AI assists, humans decide.