Compliance Guide
Financial Industry Regulatory Authority
FINRA regulates broker-dealers and their AI systems in US securities markets, with specific guidance on algorithmic trading, customer communications, and supervisory obligations for AI tools.
Agent-specific requirements
- Rule 3110: Supervisory system requirements for AI-assisted trading and recommendations
- Rule 2210: Fair and balanced communications when AI generates customer-facing content
- Rule 3120: Written supervisory procedures covering AI systems and their outputs
- Algorithmic trading controls: pre-trade risk checks, kill switches, and testing
- Books and records requirements for AI-generated trade recommendations
- Suitability obligations (Reg BI) when AI agents make investment recommendations
How Signet scoring maps to FINRA
Signet's Financial dimension directly evaluates trading agent trustworthiness. The Quality dimension tracks accuracy of AI-generated recommendations. Reliability measures uptime critical for market-hours operations. The audit trail provides the books and records that FINRA examinations require. Score decay on configuration changes alerts supervisors to review updated AI behavior.
Implementation guidance
Broker-dealers deploying AI agents should require Financial dimension scores above 85 and overall Signet Scores above 800. Configure Signet alerts for any score decay events to trigger supervisory review per Rule 3110. Use transaction history as part of books and records obligations. Document Signet Score requirements in written supervisory procedures.
FINRA-ready agents
Register your agents and get compliance-mapped trust scoring for FINRA.