The regulatory boundaries every agent and agency should know before deploying AI.
TL;DR
12+ US states issued AI-in-insurance bulletins in 2025–2026 requiring disclosure and human review for adverse actions.
Key Data Points
- 12+ US states issued AI-in-insurance bulletins in 2025–2026 requiring disclosure and human review for adverse actions.
CMS 48-hour SOA rule for Medicare Advantage and PDP. TCPA prior-express-written-consent for automated calls and SMS. HIPAA for any PHI touched by AI.
The Short Answer
12+ US states issued AI-in-insurance bulletins in 2025–2026 requiring disclosure and human review for adverse actions.
Federal & CMS
CMS 48-hour SOA rule for Medicare Advantage and PDP. TCPA prior-express-written-consent for automated calls and SMS. HIPAA for any PHI touched by AI.
State DOI
At least 12 states have issued AI-in-insurance bulletins in 2025–2026. Common themes: no unfair discrimination in rating models, disclosure when AI is used in claims decisions, and human review requirement for adverse actions.
Bottom Line
AI in insurance is now table-stakes for competitive agencies. The question is not whether to adopt it, but whether to buy an insurance-native platform (unLocked CRM ships AI tools trained on carrier data, commission feeds, and state DOI rules) or stitch together five generic tools that were built for other industries.