Login
ISCOVER
The major bank has revealed that it is currently working to automate credit decisioning in mortgages by utilising artificial intelligence and machine learning.
Speaking in ANZ’s bluenotes podcast, the chief executive officer of the big four bank, Shayne Elliott, asked Jason Humphrey, the bank’s chief risk officer (CRO) for Australia, how machine learning was being utilised at the bank.
To continue reading the rest of this article, please log in.
Looking for more benefits? Become a Premium Member.
Create free account to get unlimited news articles and more!
Looking for more benefits? Become a Premium Member.
According to the CRO, the major bank is currently undertaking a project to harness AI and machine learning to automate the credit decisioning mechanism.
He explained that while machine learning (i.e. the ability of a machine to “detect patterns in data [and] then make predictions and recommendations on how that data could and should be used”) has been used by the bank for the past 20 years, technology is now catching up to be able to process and analyse more data more quickly.
“Machine learning has been [used at] ANZ for 20 years in risk management, in the construct of what we call application scoring; we take profiles of customers [and] we look at how they perform over 18-24 months. We then determine who’s a good payer and who’s not such a great payer. We then go back and look at the attributes and build statistical models around what are the attributes that help predict that performance and that outcome. So, that’s a pretty good example of machine learning,” he said.
“Some of the techniques [e.g. neural networks, random forests etc] have been around for 20 or 30 years... but because of the complexity of those models, and the position where they are used in an ecosystem, we’ve never had the compute power to be able to run them, say in the construct of real-time decisioning.”
“Now we do, which is a revolution in itself,” he said.
The ANZ CRO outlined that Emma Grey, ANZ’s new group exec, data and automation, and John Campbell, ANZ’s general manager, home loans, were now working on “automating the home loan process”, which he said was “really exciting”.
“[We’re] looking to automate the home loan process – which is very document-driven – [and] trying to condense that, trying to extract data they can send into our decision systems for me to make a decision,” he explained.
“You’ve got the turning of documents into data… so how do I not get the documents [from the customer] but still get the data to speed up that process and end up with the same veracity of information so I can make a decision?
“For me, the really exciting part is that, in today’s world using old-school techniques, we can make a decision after all those processes have been conducted within four seconds. Under these new techniques and the new collection of data, working with our business partners, we think we can increase that percentage to a substantially higher rate,” he said.
“So, you don’t need to go and get these documents, you don’t need to do anything else, and you could get a decision sub-four seconds,” he explained.
The bank outlined that this would not only improve the customer experience (as it would reduce the need for customers to collect their documents and for the bank to manually check them), but also speed up credit decisioning.
Turnaround times in focus
The turnaround times at the banks have been a major pain point in the mortgages space, as time-to-decision has blown out as mortgage applications reach record levels. Indeed, last year there were particularly long delays at ANZ, with CEO Shayne Elliott revealing that the bank was receiving $1.2 billion in mortgage applications a day at the peak of the pandemic.
According to the most recent Broker Pulse survey (covering March 2020), brokers are reporting that ANZ’s turnarounds are at 17 days.
However, there have been concerns that banks are prioritising direct-channel mortgage speeds with a growing disparity between direct and third-party turnaround times.
For example, according to CBA’s most recent financial results, lending decisions are now automated for approximately 65 per cent of home loans coming through the proprietary channel, often for the same day. However, findings from Momentum Intelligence’s Broker Pulse survey found that brokers were experiencing an average wait of 18 days for CBA to reach an initial credit decision.
Indeed, in February of this year, the CEO of the Commonwealth Bank of Australia (CBA), Matt Comyn, told The Adviser that it’s “easier” to improve speed via the proprietary channel than broker “because of the richness of the data” that the bank has about its customers.
Some lenders have previously suggested that rework from broker applications was “a very significant factor” to longer turnarounds, while “some say that there’s reluctance of brokers to consider supporting other lenders”.
However, the Mortgage & Finance Association of Australia said earlier this year that it was “convinced” that channel conflict was “alive and well”.
The broker associations and aggregators have reportedly been working with lender heads to improve turnaround times for the broker channel.
You can listen to ANZ’s bluenotes podcast, below:
[Related: Digitising mortgages]
JOIN THE DISCUSSION