Banks Aim AI At Credit Risk, Payments Services

Of the seemingly inexhaustible uses of artificial intelligence (AI) in the financial sector, its applications around managing credit risk and optimizing payment services are among the most promising. The proliferation of “smart agents” that handle these tasks is a glimpse of more innovation to come, as AI proves its worth to financial institutions (FIs) in the great reopening.

That FIs and enterprises are pouring millions into AI development is not surprising, and where they’ve been focusing that development shines a light on where AI can do the most good.

The April 2020 Unlocking AI Playbook: Credit Risk And Payments edition, done in collaboration with Brighterion, notes that “… banks appear to be applying AI with greater specificity than in the past, suggesting that strong use cases have emerged. Our research shows that 92.9 percent of AI-using FIs are applying it to payment services, and 71.4 percent are doing so in credit underwriting. The latter finding is a marked shift from our 2018 study, in which just 27.3 percent of FIs reported using AI in credit underwriting.”

Decisions, Decisions

FIs and their tech partners have been figuring out how best to deploy AI for the past few years, and now real-world applications that move the needle on problems like credit risk are in the field, and the outlook among banks has shifted to a better understanding of what AI does well.

“An overwhelming majority of the banks that currently use AI apply the system in payment services: 92.9 percent,” the latest AI Playbook states. “The second-most common application is credit underwriting, as 71.4 percent of AI-using respondents deploy it in this area.”

While banks are putting AI to use chiefly in payment services for security-related functions like system security (69.2 percent), authentication (61.5 percent) and preventing data breaches (53.8 percent), “… the most common use of AI within managing credit risk is in credit decisioning, which was cited by 60 percent of AI users — more than twice the share citing any other related activity,” the report states.

Credit risk decisioning is a breakout application for AI, with many banks reporting that the technology is dramatically streamlining that set of tasks.

“… 60 percent of AI users reported deploying the system for credit decisioning,” the latest AI Playbook states. “AI is making in-roads into credit/loan underwriting as a way to supplement conventional scoring models — or supplant them altogether. In this way, AI is being used to expand the pool of potential borrowers — and to bring greater scrutiny to potentially risky ones who might only look good on paper. An array of FinTechs are offering services in this area.”

Payment Services Get an AI Assist

AI-powered smart agents hold immense promise for managing payments, with security implications that represent a gamechanger for some organizations. For example, FIs could deploy a smart agent for every single card they issue, bringing a level of control program managers didn’t dare dream about until fairly recently.

“Our research shows that 43.7 percent of FIs that are at least ‘somewhat’ interested in using smart agent-based AI would be ‘very’ or ‘extremely’ interested in using it for their payments business units,” according to the April 2020 AI Playbook.

“This is followed by their new product or customer life cycle management units (38.8 percent) and credit risk underwriting units (33 percent). The concentration of interest in these three units represents a marked shift from 2018, when the largest share of FIs (50.5 percent) were most interested in using smart agent-based AI to address internal fraud.”