What should the AI section of an RIA’s SEC exam file contain about Claude?
The AI section of an RIA’s SEC exam file needs six documents: the written AI usage policy with a named supervisor; vendor due diligence for Anthropic (Trust Center certifications, commercial terms, DPA, subprocessor list, each dated); Reg S-P service-provider oversight notes and the incident response plan showing where Anthropic’s 48-hour notice lands; supervision evidence (the monthly admin review log); the marketing-rule check if Claude touches client communications; and the training record by person. Strategized.ai’s Claude implementation builds all six during the rollout so the exam request list is a folder, not a scramble.

In one picture
Diagram: Six documents for the exam file
What Anthropic and the rules actually publish
The follow-up questions owners ask
Will SEC examiners ask to see our Claude prompts?
They can ask how the firm supervises use and whether records exist. On Team, the review log and retention rule are the record; on Enterprise, the Compliance API can produce sessions. Decide which your program promises before the exam.
Is client data in Claude a Reg S-P issue for an RIA?
It is a service-provider oversight item: contract, due diligence, breach notice path. The DPA and the Trust Center are the oversight artifacts.
How should an RIA’s SEC exam file treat Claude or other AI used in marketing?
If Claude drafts client-facing material, the marketing rule review applies as it would to any draft. Keep the review evidence.
What this means for your rollout
Related services and tools
Get the plan before you buy the seats
The plan choice is the easy part. Configuration, governance and training are where firms get it wrong. The Claude Build Plan is two working sessions with Michael Robichaud, then a written plan: architecture, plan choice with the reasons, your firm’s custom Claude plugin scoped, rollout timeline and cost. A fixed , and the plan is yours to keep.
Related questions
Sources
- SEC Regulation S-P amendments: https://www.sec.gov/rules-regulations/2024/05/s7-05-23
- Anthropic DPA: https://www.anthropic.com/legal/data-processing-addendum
- Anthropic Trust Center: https://trust.anthropic.com/
Strategized.ai is independent and not affiliated with or endorsed by Anthropic. Claude is a trademark of Anthropic, PBC.