Banks that use generative AI to answer customers now face a question from federal examiners that many of those systems struggle to answer: Where did that answer come from? That is the central finding of an analysis Inbenta published today in its white paper Closing the Promise Gap: How Encore Supports the OCC Supervisory Framework for AI. The analysis examines the supervisory framework for AI in banking, recently issued by the Office of the Comptroller of the Currency (OCC).
The OCC framework is the first AI-specific examiner framework from a federal banking agency. The framework warns that AI hallucinations about fees, rates, account status, or policies could create legal exposure and customer harm. And it makes clear that the bank, not just the vendor, is accountable.
Inbenta’s analysis finds those concerns converge on customer service, which is the source of the greatest exposure for banks. The Deloitte Consumer Banking Survey 2025 found that 58% of respondents had used a bank’s chatbot. When a generative model writes each answer live as the customer waits, the risks follow: a wrong fee, an answer that shifts from one request to the next, and a thin audit trail.
Read More on Fintech : Global Fintech Interview with Andy Jiang, Head of Product, Slash
“Many banks haven’t read this framework yet, and many are running customer-facing AI that writes answers on the fly,” said Melissa Solis, CEO of Inbenta. “Sooner or later, that design was going to collide with an examiner’s first question: Where did this answer come from?
Source link







