AI works first on clear rules, and banks must explain loans

Agents will first handle simple tasks with clear rules, while a law firm says that in many cases nobody at the bank formally signed off on AI in lending.

By @0xNotMarc

Rule-bound work goes first

A UOB report says AI agents will first handle simple, repetitive tasks with clear rules, before moving into more complex financial decisions. NatWest has brought in AI agents from Adclear to overhaul how compliance is embedded across its marketing and communications operations, slashing approval times by 80%, according to Finextra.

Lending is drifting too

Outside GC, a firm that provides outside legal counsel to companies, warns that AI tools now score borrowers, sort through financial statements and write first drafts of loan approval memos. In many cases, it says, nobody at the bank formally signed off on that change.

Explaining a decision

The same article says banks must be able to explain loan decisions to regulators, borrowers and other lenders. Its advice: keep an up-to-date map of where AI shows up in the lending process, and when a person overrides the tool, record it and explain why.

Customers hold the log-in

On the customer side, Bank of America emailed customers to ask them to never share their username and password with an outside service, including an AI-enabled tool or agent. Account agreements at most banks hold customers responsible for anyone they give their log-in to, according to TechCrunch.

A map to draw this week

Whatever role we hold in a bank or lender, a useful exercise is to list where an AI tool already touches our own files, and who signed off on it. Would we be able to show a regulator or a borrower where a person stepped in?