What you need to know:

• Seven years in the making, Commbank's customer engagement engine is moving the bank beyond traditional banking services.
• Its Benefits Finder services has put an average of $710 into the pockets of customers utilising the app, facilitating payment of more than a billion in grants into communities.
• Its AI-powered Smart Weather model means the bank can contact customers ahead of damaging weather events so they can prepare, and it also makes it easier for the bank to provide remedial support in a crisis.
• The engine, built on technology provided by Pega, makes 53 million calls a day just on inbound activity, and is enabling the bank to run thousands of experiments in market weekly.
• Rachel Hansell, Executive Manager, Personalisation, Commonwealth Bank of Australia, says the bank could do more and go faster if it wanted to.
• Suncorp, Optus, TPG, Foxtel, ANZ Plus, AGL, Toyota, and BUPA have built real decisioning engines, Mi3 has been told, while NAB signed with Pega last year, and is now believed to be planning its implementation.

Commbank has been deep on AI for years. Three years ago data & analytics chief Andrew McMullan told Mi3 that its intelligence and decisioning power had changed the way the bank does business – putting millions back into customers pockets as the pandemic hit, and creating ten times more homeloans leads that were 300 per cent better than anything they had seen before while helping to solve in-branch and digital disconnects.

But that was just for starters.

Now Commbank is using AI-driven decisioning to help deliver a new generation of apps beyond the traditional domain of banking services, designed to reflect a customer-obsession culture that breaks down and replaces the product fiefdoms so characteristic of banking around the world. But the fundamental aims remain the same: put more money into the pockets of its customers, help build loyalty and ultimately life time customer value.

It's not the only corporate taking that approach. The real time decisioning model is particularly attractive in the financial services sector and also in telco where companies have massive scale – revenues in the billions, millions of customers, and a broad range of products and services.

In Australia, in addition to CBA, companies like BUPA, which is achieving impressive response rates and strong NPS scores, Suncorp, Optus, TPG and Foxtel have implemented or are implementing decisioning engine models. ANZ takes that approach within its digital division for ANZ Plus, but not yet in core banking. NAB meanwhile is believed to have signed with Pega last September. AGL and Toyota are also believed to have implementations.

But Commbank remains ahead of the pack. Along with Dutch Rabo Bank, it's considered one of the top real-time decisioning implementations in the world and is believed to have as many as 15 instances of Pega beyond just its Customer Engagement Engine, in areas such as credit risk decisioning and intelligence applications. It is also using the technology within external advertising, plugging in its first party data via the decisioning engine to the Google ad stack so that it can stop targeting customers that are not suitable for certain offers, and better communicate to those that are while also future-proofing against the deprecation of cookies.

Diverse CX

In a recent presentation at Pegaworld in the US, Rachel Hansell, Executive manager, Personalisation, Commonwealth Bank of Australia, described an approach to personalisation and customer experience extending far beyond the typical next best action campaigns – which are too often just code for next best sale, according to critics, and instead are designed to create much richer, multi dimensional customer relationships that deliver obvious, immediate and tangible benefits to the bank's customers.

It is now using some of that same technology to reinvent what banking means in a world of digital connectivity where brands compete on customer experiences.

Hansell’s job is to bring that promise to life using data analytics, AI and information technology to deliver highly personalised experiences. "We work with business people, digital people, data scientists, and behavioural scientists, and think that diversity is what helps us to do different things, because of the creativity and thinking comes from so many different people.”

The customer engagement executive outlined a range of applications the bank has been able to build as of a result of its investment in the customer engagement engine. The benefits finder tool, which makes it easy for the bank's customers to determine what government payments they are entitled to, and to actually get the money, has already unlocked more than a billion dollars of grants, giving customers that have used the tool an average of $710 apiece.

“It’s actually quite difficult to find some of these benefits, so for customers that are digitally active with us, we set them up for ID verification, to let them know what benefits are available," Hansell explained to delegates. "That takes [them] through to our hub where we use machine learning to decide which benefit is best for the customer. We use machine learning to personalise and rank the order in which we show those benefits. The customers can then also explore the other benefits that are available, and there's hundreds of them. Using the customer engagement engine allows us to proactively contact customers about what benefits are available to them outside of the bank, and it puts money back into their household budget.”  

New front

Hansell also outlined what the bank describes as its smart weather model, which makes it faster and easier to provide banking and finance services to customers who find themselves on the wrong side of Australia’s notoriously unfriendly weather conditions, be that bushfires or floods from torrential storms.

The service builds on that spun up in 2019 to help customers affected by the worst bushfire season in living memory, where the bank hooked up to fire service feeds to understand by postcode where fires were burning, and where they may be headed, to contact customers and tell them what help was available.

“We know that these [weather events] come out of the blue and we can’t prevent them. But we really wanted to do something to help the wellbeing of our customers when they are facing adverse weather events,” she said.

So starting in 2021, the digital science team began developing a purpose built AI-driven smart weather model which pulls data from sources across Australia, including emergency services, along with the bank’s own internal insurance information, to continuously monitor for adverse weather events. 

This enables the bank to identify and contact customers who are effected by adverse weather events, and offer them tailored support packages. 

“If you are a customer and have a home loan with us, if you're in a region that has been impacted by an adverse weather event we will offer you a home loan payments deferral. We have other packages for business customers,” she said.

The packages are tailored to particular circumstances. “We’re able to respond in hours to natural disasters, and we feel we can credibly respond to multiple natural disasters at once," said Hansell. "We also have a small team in the bank that are constantly monitoring to identify events before they happen so that we can go out to customers before the event happens and give them those few hours to prepare for what is coming.”

She said there have been instances where Commbank customers were contacted by the bank before they heard from emergence services. 

Massive scale

Now into its seventh year, the bank's Customer Engagement Engine was first deployed in the branch networks – starting from a single branch in Penrith. Now it is at the heart of its retail operation. CEE makes 53 million calls per day just on the bank’s inbound platform. “If I included outbound it would be a lot more,” said Hansell.  It processes a hundred to several hundred million data points across 1,000 machine learning models each day.

At the moment the engagement engine is making decisions to distribute real time notifications and alerts, customer service messages, engagement communications, marketing communications and offers, and it does that across all channels – mobile apps, online banking, call centres, branches, SMS, email, chat, and ATMs she said.

It has also given the bank a powerful platform for experimentation. At any given time there can be "thousands" of those experiments in market, per Hansell. 

"So it's quite scaled, and we release optimisations and experiments twice per week. We can do more if we want.”