Customers may not realise AI algorithms are doing some of the work behind the scenes when they deal with a bank but it is happening, as lenders compete to offer better or more-efficient services.

“In the future, you could expect solutions that help people understand their money better and help manage their financial wellbeing.”

Tim Hogarth, ANZ’s chief technology officer

Mortgages are a case in point. Westpac’s Walker explains that the bank has been piloting an AI system to help filter home loan applications that it receives through mortgage brokers, to give brokers early feedback.

“What we’ve done is we’ve combined different forms of artificial intelligence to triage the mortgage application and to assess it against things like our policies, our product statements, the way that our products are defined and, obviously, completeness in terms of information,” Walker says, adding that there will be ways to use AI elsewhere in mortgage processing.

National Australia Bank, the country’s biggest business lender, has also started using AI to automate the review of legal documents that must happen when customers want to borrow through a trust. It currently takes three days to give these customers a decision: the bank hopes to get this down to one day by using AI. This is just one of 20 AI “uses cases” NAB is working on.

It is also likely banks will use AI to soup up their apps, as lenders compete to give their customers useful automated financial tips and advice.

ANZ’s chief technology officer Tim Hogarth says that in the future: “You could expect solutions that help people understand their money better and help manage their financial wellbeing, so they can get a more personalised response through perhaps a digital coach.”

AI can also help the banks pick up risks, such as customer facing financial difficulties. ANZ, for example, says it has an AI system that can identify bank customers at risk of financial distress about 40 days earlier than usual, based on transaction data.

Scam detection is another area ripe for AI assistance. With Australians losing $2.7 billion to scams last year, banks say AI can help them spot cases where customers may have been tricked into sending their money to scammers (who are also using AI in many cases).

Westpac’s Walker says that when you make a payment, AI helps to assess all sorts of variables, including how much money is being sent, whether the money is moving overseas, and what you’ve written as the reason for the payment. If a pattern emerges that is similar to previous scams, the bank will alert the customer to this risk. “We’re having to keep one step ahead of [scammers] all the time, adjusting what we do,” Walker says.

Loading

Behind the scenes, thousands of software engineers inside Australian banks are also already using AI to help do their jobs.

ANZ’s Hogarth says the bank has 3000 engineers using AI, and testing has suggested there were cost savings of between 40 per cent and 55 per cent on using the technology for some specific programming tasks – though he added engineers don’t spend all their time writing code. He says these are “quite material” savings.

All of which raises the question: if banks are already using AI significantly, and they plan to do so even more, will the technology replace some human workers?

This has long been a worry for the Finance Sector Union, which is concerned about the growing adoption of AI in finance, and what it says is a lack of transparency about how the technology is being used.

Loading

FSU national secretary Julia Angrisano says AI is being used in areas such as rostering of shifts, and the monitoring and management of some members’ work.

“AI can certainly increase the efficiency of work, but we are seeing those benefits only going to employers in the form of ever-growing profits, with none of these benefits flowing down to workers,” she says. “In fact, many major employers in the finance sector are being sold on AI programs for the purposes of cost-cutting and thus eliminating vital jobs.”

Bank executives are often vague when discussing what AI might mean for staffing in the longer-term, but for now, they say that while some tasks can be completed using AI algorithms, the technology is also creating new roles. Comyn, for example, acknowledges some simpler engineering tasks can probably be done by AI, but he says the bank’s broad goal is to get more done “with the same broad investment allocation that we’ve got”.