Agentic AI in transaction monitoring: How banks can scale AI in compliance
By McKinsey & Company ·
McKinsey shows how one transaction monitoring use case can become the base for scaling agentic AI in compliance, drawing on a proof of concept at a European bank. Banks have cut human processing time in investigations by more than 40 percent, with 30 to 50 percent of the agent network dedicated to quality control and traceability, and a redesigned process likely needs 15 to 20 squads of three to five agents.
Key figures
| Figure | Value | Source |
|---|---|---|
| Reduction in human processing time, transaction monitoring investigations | more than 40 percent | Web page |
| Share of the agent network dedicated to quality control and traceability | 30 to 50 percent | Web page |
| Agent squads needed for a redesigned process | likely 15 to 20 squads of three to five agents each | Web page |
| Fall in process effort with agentic AI | 30 to 50 percent | Web page |
| Proof of concept scope | customers with more than 15,000 transactions; the same eight web-based checks on every case | Web page |
- Global