FICO urges UK lenders to use AI in collections for payments

FICO has urged UK lenders to use artificial intelligence in collections operations, following its analysis of UK credit card data for 2025.

The share of customers missing two or three card payments rose through much of the year, while balances on those accounts also increased from a year earlier. Accounts with two missed payments reached an average balance of £2,938 in November, up 4.9% year on year. Accounts with three missed payments climbed to £3,324 in December, up 4.1%.

The trend is adding pressure to collections teams already dealing with larger case volumes and more complex customer circumstances. Many lenders still rely on manual handling, with staff reviewing cases one by one, a model FICO described as expensive and difficult to scale.

It argued that this has left many operations dependent on older tools and processes that do not adapt well when payment problems rise. FICO highlighted scripted systems such as diallers and basic chatbots, saying they can improve productivity but are often limited to simple interactions.

Mike Trkay, Chief Information Officer at FICO, said lenders face a growing mismatch between demand and the way collections work is organised.

"Collections is one of the clearest examples of where traditional processes are no longer fit for purpose," Trkay said.

He said lenders are trying to manage high volumes, limited resources and changing customer expectations at the same time. In FICO's view, conversational AI can support more natural exchanges with borrowers and gather information as discussions develop.

Operational strain

One area of focus is how AI systems identify signs of customer stress and vulnerability. Those signs are not always obvious and may differ between people with a long history of paying on time and those who fall behind more regularly.

A borrower who misses a payment because of affordability pressure may need a different response from a customer with a repeated pattern of late payment. In that context, FICO said AI tools could help direct cases to more appropriate treatment and reduce unnecessary manual work.

At the same time, Trkay said conversational systems create risks of their own if lenders do not tightly define the rules they operate under. Because these systems can follow a customer conversation in many directions, they increase the chance of inconsistent outcomes.

"Basically, the AI needs to know when it should stop and make a hand-off to a live agent," Trkay said.

That point is likely to resonate with lenders navigating consumer protection obligations and internal conduct rules. FICO said AI systems used in collections should be limited in the language and guidance they provide, and should recognise when an interaction is edging into advice or becoming too prescriptive.

Decision tools

Beyond conversation handling, FICO said the bigger opportunity lies in combining those systems with decision intelligence. In practice, that means using AI not only to understand what a customer is saying, but also to choose an action based on policy, affordability signals and expected outcomes.

That could support automated negotiation of payment plans, identify customers in financial hardship, guide borrowers through repayment options and keep interactions within regulatory and internal policy limits. AI-driven optimisation could also help lenders decide where to deploy limited staff and resources across large portfolios.

This would involve weighing the cost, effort and likely returns of different interventions, rather than applying the same collections treatment across broad groups of accounts. FICO said that could help lenders prioritise cases where engagement is more likely to lead to recovery or where a more tailored approach is needed.

Trkay said the aim is to move away from standard responses that fail to reflect differences between borrowers.

"AI enables organisations to move beyond static, one-size-fits-all approaches," Trkay said. "It allows lenders to deliver more personalised, responsive and effective collections strategies, improving outcomes for both the business and the customer."

The comments come as lenders continue to monitor the effect of household financial pressure on unsecured borrowing performance. Rising missed payments on credit cards can point to broader strain in consumer finances, while also increasing servicing costs for banks and card issuers if account management becomes more labour-intensive.

FICO warned that lenders that do not update their collections approach risk falling behind as volumes change and operating pressure grows. "Organisations that leverage AI effectively will not only improve efficiency but also build stronger, more resilient customer relationships," Trkay said.