CBA Calls Machine Learning Models After Avengers – Finance – Strategy – Software

Machine Learning

Australia’s Commonwealth Bank named its machine learning models after superheroes and makes them “compete against each other every day” to deliver the best customer service results.

Through its partnership with, the bank hopes to create both superhero-like models and model builders, accelerating its use of machine learning.

Senior Product Owner Avinav Goel said at yesterday’s webinar that the “human side” of the program was “easily [his] Favorites section.

He said banks have “literally gamified the art of modeling between analysts and decision-making scientists.”

“We told them, ‘You’re not building a propensity model, you’re building ‘superheroes’ who will compete with each other every day to better serve customers.'” Goel says.

“The benefits model, Captain America, led the charts in week one, the business model, Thor, beat him in week two, and Iron Man came off a dominating week one. A start, two, but by week three he was down and told us, ‘Hey, this needs a model rescore.'”

By gamifying the program, more insights were created and shared by the team, also creating a “community of learners within the bank across geographies.”

According to Goel, CBA teams can ask questions and get answers from each other “in real time through group chats and discussion boards.”

“These fellows, if I could call them [that]and their “superheroes” [models] Exists in CBA’s proud universe. We call it the ‘Superverse,'” he said.

Scaling up CEE

One of the bank’s earliest AI/ML efforts was the Customer Engagement Engine (CEE). It suggests sub-optimal conversations with each bank customer.

CEE now has “enough science, business rules, and adaptive models to personalize” conversations, but Goel said banks wanted to future-proof it, but they didn’t want to. He said he was acutely aware of the amount of resources needed to

This has resulted in the adoption of the platform, and significant upskilling efforts are currently underway across the bank.

“We also knew that, done the right way, this would bring analytics closer to the rest of the organization,” says Goel.

According to Goel, the initial integration into CEE will be on a small scale with the goal of increasing the percentage of customers who engage with conversations, offers, or other actionable items placed in front of them. and started experimentally.

Prior to the change, CEE was producing “better customer outcomes overall.”

However, after introducing the new model, “two out of three customers in the H2O segment were encouraged to take action, highlighting the success of using as an avenue to improve CEE.

Banks will see similar engagement improvements in other existing tools by enhancing them with built-in models, such as the “Benefits Finder” government rebate search tool, the “Saving Habits” tool, and the “Billing Sense” payment prediction product. said to have seen

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *