Banking on AI: How a new operating model can transform distressed credit management
By McKinsey & Company ·
McKinsey argues that banks using AI in collections and distressed credit should move from separate use cases to an agentic operating model that coordinates segmentation, outreach, negotiation and restructuring. The pressure is building: US serious delinquency flows remain high, debt-management complaints to the CFPB passed 200,000 in 2024, and borrowers offered a practical digital route are 20 to 30 percent more likely to pay.
Key figures
| Figure | Value | Source |
|---|---|---|
| Organizations scaling AI across the enterprise / fully deployed | fewer than one in three / 7 percent | p. 1 |
| US flow into 90+ day delinquency, Q2 2026 (annualized) | automobile loans 3 percent, credit cards 7 percent | p. 2 |
| Brazil household delinquency on nonearmarked credit | up more than 40 percent over 18 months to a record 8 percent | p. 2 |
| CFPB debt-management complaints | more than 200,000 in 2024, up from approximately 110,000 in 2023 | p. 2 |
| Repayment with a practical digital route vs outbound calling alone | 20 to 30 percent more likely to pay | p. 2 |
- Global