Banks switch on AI as customer battleground intensifies and profit margins fall

Financial services are in a race to deploy artificial intelligence in almost all facets of their operations to win more customers.

Bank of Queensland chief information officer Craig Ryman says AI allowed businesses to compete in that they couldn’t previously.

Artificial intelligence is being deployed across Australia’s financial system to speed up money transfers and make it easier to open bank accounts as the much-hyped technology becomes the new battleground to win market share.

NAB and Bank of Queensland have tapped into Microsoft’s Copilot AI platform to lift productivity by removing tedious and time-consuming tasks, such as coding, writing reports and drafting job advertisements.

Airwallex – a self-branded challenger to the big banks, valued at $US5.6bn – is using AI to vet new customers more quickly as part of its “know your customer” checks, accelerating its global growth.

NAB chief executive Ross McEwan.

NAB chief executive Ross McEwan has hand-picked a tech team to design a powerful AI model which can sift through billions of real-time interactions with customers to glean a better view of their habits and anticipate their needs.

ANZ is building AI features in its new banking platform ANZ Plus, which has begun offering digital home loans.

Intense competition among the big banks has crunched profit margins, prompting many to turn to technology in an effort to slash their cost base, while winning over new customers.

But they say the widespread adoption of AI isn’t about to spark a mass staff cull, despite computers and machines increasingly assuming more human tasks.

Bank of Queensland chief information officer Craig Ryman said those concerns were based on the assumption that as productivity lifts, output stays the same. Bank of Queensland engineers are using Github Copilot – owned by Microsoft – to write better code, which they are now able to do in half the time.

“The reality of that will be that to do the same unit of work, you need less engineers,” Mr Ryman said.

“But in fin services, and what I’ve got experience in, there isn’t a lack of work to be done. It’s more that you prioritise, and you get done what you can get done with the capacity that you have.

“I’ve got a sort of philosophical position on it, which is by freeing up productivity, it’s actually allowing businesses to compete in areas that they don’t have capacity to compete in.

“We have a long list of things that we would like to digitise more for our customers or make it easier for our people to operate inside BOQ and it’s effectively going to speed up our ability to realise those opportunities.”

Bank of Queensland has also AI to produce a procedure document in about 45 minutes – a process which previously required a three-week turnaround. It is also using AI to write job advertisements and interview guides, fastracking recruitment.

At Airwallex – which is recruiting 500 new staff – it has halved the number of “false positives” in “know your customer” (KYC) checks.

Previously, Airwallex used rules-based analytics and natural language processing to scan new customers’ websites. Those tools are very effective in identifying high-risk keywords, but also tend to generate a significant volume of false-positive alerts – which can slow down the KYC process unnecessarily.

For example, previously a business selling military-style jackets would attract greater scrutiny because it would be confused with selling military equipment. This would lead to a staff member having to then physically vet the business – a process which could take days, as was the case when Silicon Valley Bank collapsed earlier this year and customers tried to switch banks.

Airwallex global head of product Shannon Scott.

But Airwallex global head of product Shannon Scott said the new AI platform had boosted the number of customers who get onboarded, without human intervention, by 20 per cent.

“It’s something that is happening in the background. But when you can actually get that customer from sign-up, and just a few minutes later – with no manual intervention – they’re up and running, it’s a really great customer experience,” Mr Scott said.

“It’s what the customers expect but it’s not necessarily the reality when regulations in different countries require us to go through a lot of different processes to ensure this is a suitable customer for onboarding.

“Many other financial institutions are doing this over a week. We can do it, very often, now within minutes.”

Airwallex co-founder and chief technology officer Jacob Dai said the company was building a suite of new services, which it will launch next year, using generative AI and natural language processing (NLP) to enhance customer experience and win market share.

“We know our customers want tailored services, immediate assistance, and intuitive access to their data,” Mr Dai said.

“Using generative AI and NLP, we can help our customers get answers quickly and effectively, whether it’s searching for a specific transaction in their history, or creating a detailed payout workflow with an automated chat function. There are so many applications to explore, and we’re just getting started.”