ITnews

Westpac plugs five AWS AI agents into core lending processes


Westpac is using a series of AI agents built on Amazon Bedrock AgentCore to process payslips provided by customers when applying for a home loan or credit card.



The bank revealed in June that it is embedding AI into core transactional “flows”. It is now clearer what this looks like in its home lending and consumer finance domains.

Chief data, digital and AI officer Andrew McMullan told the AWS Financial Services Symposium in Sydney that AI has become “a core part of how the organization delivers value to our customers”.

“We have been partnering with AWS to use [Amazon Bedrock] AgentCore to remove the complexity from some of our lending processes,” McMullan said.

“We’ve now got specialist agents working alongside our bankers to make our customers’ lives easier.”

McMullan said that every week, the bank processed over 32,000 payslips and “more than 1.5 million transactions” as part of its analysis of income and spending habits when customers make loan or credit applications.

“For customers, a payslip is just one document in an application process. For our people, verifying it can be really complex, time-consuming, and full of policy judgements,” McMullan said.

“Now our specialist agents work together around the banker. They can classify the payslip, they read it, they can extract the data, run the calculations in the back, and verify it against policy before giving our banker the output to review and continue the application with speed, confidence, and trust. 

“This now means our bankers spend less time wrestling with documents and much more time helping our customers move through one of the biggest financial decisions of their lives.”

McMullan said that payslip analysis is providing similar benefits to the credit card application process, primarily helping to avoid delays.

“Customers can sometimes upload the wrong document – maybe the payslip’s a little bit out of date,” McMullan said.

“Historically, that would have created a delay, [or] some rework for the banker. 

“Now our bankers are working alongside the specialist [AI] agents, and we provide feedback in real time, so customers can correct the issue immediately and continue with confidence. 

“For every customer, every day matters. Avoiding delays matters,” he said.

McMullan was clear that the agents are “running in [Westpac’s] production systems, and are not merely prototypes.

“We have specialist agents that are extracting the data, applying policy, and validating outcomes alongside our bankers,” he said.

“Our bankers continue to have oversight and help customers make decisions. For us at Westpac, it’s all about removing friction and complexity from the process, so our customers who are applying for a mortgage or waiting for an approval for lending are getting to an answer more quickly and with greater confidence.”

The bank showed a slide with five agents running in its lending and credit application processes.

Within mortgage applications, an “account scrutiny agent” is used to process manual bank statements and classify transactions and spending, while two additional agents “extract payslip data [to] speed up income assessments” and “verify payslip details against policy and required calculations”.

The two payslip agents are also used in the processing of credit card applications by consumers.

McMullan said that the assistance of the AI agents in lending processes produced “more than 150,000 hours” of time being saved, which could be “reinvested” by teams to serve customers better.”

He added that the bank has reached a stage of maturity in its agentic work with AWS where it has shrunk the time to get an agent in production from an initial six months, to “about four-to-six weeks”.

“In partnership [with AWS], we’re moving at pace,” he said.

“Our ambition is simple. We want to build agentic operations to support our people [to] serve our customers better. 

“Agentic operations are about reducing the distance between intent and outcome. Decisions will now move faster because intelligence is available exactly when it’s needed in the process.”

Ry Crozier attended the AWS Financial Services Symposium Sydney as a guest of AWS.



Source link