AI in credit risk: an aid to judgement

This briefing from the Risk.net Credit Risk Leaders’ Network held in London in October 2026 captures insights from senior credit risk professionals on where artificial intelligence is adding value, the data and governance challenges of scaling it, and the risks of relying on third-party vendors.

Among the takeaways:

• People still make the credit decisions: AI gathers information and drafts memos, particularly in wholesale credit
• Experimentation has produced noise as well as gains: banks are increasing use case inventories and reining in inconsistent tools
• Poor data limits what AI can do: many AI projects are really data clean-up exercises
• Independent challenge is under strain: AI drafting and AI reviewing the same work raises questions for regulators
• Participants were wary of vendor lock-in: internal tools are catching up, and dependence brings lineage, cost and concentration risks
• Accountability stays with people: banks must also develop the next generation able to spot AI’s mistakes

Download the report to explore how credit risk leaders are putting AI to work while keeping clear lines of accountability.