AI improves CommBank's customer engagement up to 400 percent

The bank used AI to analyse 200 billion data points to free up more time for its humans to do human things, like exhibit empathy and understand a customer's individual circumstances.
Written by Stilgherrian , Contributor

Artificial intelligence (AI) has helped Commonwealth Bank increase customer engagement with some routine communications up to 400 percent, according to the bank's head of data science, Dan Jermyn.

The bank has created what it calls the Customer Engagement Engine (CEE), its "flagship implementation" of AI. It's a "centralised decisioning platform" that helps "orchestrate the customer experience".

"We want to make sure we're talking to them about the right things. That's incredibly difficult," Jermyn told the inaugural Ethics of Data Science conference at the University of Sydney on Friday.

The bank has millions of customers who might want to talk about any one of thousands of different issues, via tens of different channels.

"You can imagine how much information we have on our customers. All the financial transaction information, all of the interactions," Jermyn said.

"The answer to what is the best thing to talk to a customer about at any point in time is there, if only we could decode it."

The CEE currently sits across 19 communications channels, and analyses 200 billion data points and "thousands of combinations" of the next best thing to talk about, delivering its decision in "a tiny fraction of a second".

Also: Artificial intelligence: Not the apocalypse, more of an invisible assistant

In some cases that information is used to inform customer service agents.

"Then the human takes over, and all of those skills that we can't replicate with machines -- the social empathy, the kind of understanding of the particular circumstances that the person in front of you, and what's important to them or not, what's their emotional status," Jermyn said.

"All we've really done is free up more time for our humans to do this thing that really matters, and to do the things that they're really good at, and everybody benefits," he said.

"It points the way to what the best kind of implementations of AI in industry should be trying to do."

See: Managing AI and ML in the enterprise (ZDNet special report) | Download the report as a PDF (TechRepublic)

In other cases it's used to drive "smart alerts".

Jermyn gave the example of one of the bank's routine communications, a smartphone notification that read: "Reminder: Your personal loan repayment's due in 7 days. Don't forget to pay to avoid a late fee."

"The biggest difference that we see of all, up to 400 percent, is when we start thinking about how we've decided to present that, to whom, and when, and in what circumstance," Jermyn said.

Commonwealth Bank is also using AI to analyse the factors that contribute to a customer's sense of financial well-being.

"The second stage in the ethical implementation of AI within an organisation is not just to do things which seem like the right things to you. It's to know that you are doing the right things, to measure them properly," Jermyn said.

See: AI and big data vs ethics: How to make sure your artificial intelligence project is heading the right way

The bank has built on existing research into how people perceive their own sense of financial well-being. Using its knowledge of customers' financial transactions and other measures, it found that reality has more extremes than people's perceptions.

While 9 percent of people reported having trouble with their financial well-being, the bank's measures put the figure at 18 percent. At the other end, while 12 percent of customers reported that they were doing great, the bank reckoned that in fact 15 percent were doing great.

We might imagine that the greatest drivers of a sense of financial well-being would be salary and income. But the bank found that the factors which most affected customers' perception were behavioural -- patterns of spending and saving, and how they managed their money.

"Behavioural patterns were everything in this," Jermyn said.

"One of the most exciting things about AI for us is that what it's actually allowing us to do is to produce better outcomes for our customers," he said.

"We couldn't do this without AI. It's literally impossible."

Related Coverage

Google's banana throwing robot is highly accurate

Machine precision enables this robot to throw objects through the air with the greatest of ease.

Tokyo 2020 Robot Project powered by Toyota, Panasonic

Robots are going to take over Tokyo during the upcoming Olympic and Paralympic Games, which will be as much technology showcase as athletic competition.

Pure Storage, Nvidia, Cisco launch AI hyperscale systems

Pure Storage is expanding its AIRI lineup of systems for artificial intelligence and machine learning workloads.

These are the biggest dangers of artificial intelligence (TechRepublic)

Automation is a good thing, says Starmind founder Pascal Kaufmann, but business might not be prepared for automation at mass scale.

Will AI lead project managers to the unemployment line? (TechRepublic)

Project managers may fear that AI will replace their jobs---when in fact it just might improve them.

Editorial standards