AI Compliance Agent
The AI Compliance Agent is the core decision-making engine within CRYMBO's AI layer. It processes compliance requests in real-time and determines the optimal compliance path for each transaction.
What the Agent Does
- Receives transaction context — Sender, receiver, amount, asset, jurisdictions, chain
- Determines applicable rules — Identifies which regulations apply (Travel Rule, MiCA, local KYC/AML)
- Assesses risk level — Evaluates transaction risk based on counterparty history, amount, and jurisdiction
- Recommends compliance path — Specifies which identity fields to request, which validators to engage, and what thresholds apply
- Outputs decision — Returns a structured compliance decision consumed by the Oracle validation flow
Decision Outputs
| Output | Description |
|---|---|
PROCEED | Transaction meets all compliance requirements |
REQUEST_IDENTITY | Identity exchange required before proceeding |
ENHANCED_DUE_DILIGENCE | Higher-risk transaction — additional verification required |
BLOCK | Transaction violates compliance rules — should not proceed |
MANUAL_REVIEW | Edge case — requires human compliance officer review |
Configuration
Institutions can configure the AI Compliance Agent to:
- Set custom risk thresholds
- Define institution-specific compliance rules
- Override AI recommendations with manual policies
- Exclude specific jurisdictions or asset types from automated processing