The tools and training transforming NAB’s workforce to become “the person with AI”

NAB user experience designer Kimberley Kam is explaining how she uses Cursor.ai - among the latest crop of artificial intelligence tools - which combines a large language model, software development environment and even the ability to connect to popular design tools.  

“This fundamentally changes what we can do,” Kam says. 

“As designers our question is always ‘are we building the right thing in the first place?’ – This lets us rapidly get to user testing to answer that.” 

In skilled hands, tools like Cursor are enabling NAB’s tech workforce to radically transform how they build new digital products and experiences. 

“Cursor dramatically accelerates our workflow, without ceding control - what I really like about it is it forces you to clarify your intent,” says Kam, explaining the accountability by design. “Everything it produces has traceability, so you're still owning everything that comes out as projects develop.”  

NAB Group Executive Digital, Data & AI Pete Steel explains NAB has identified four key areas of focus for its AI strategy.  

“We believe that by focusing on customers, colleagues, software development and operations we can provide improved customer experiences and deliver significant productivity benefits. 

“Ultimately we want to spark innovation that delights our customers – this is about enabling that creativity, being able to test, learn, develop and deliver those experiences much faster,” Steel says. 

Re-imagining software development 

Group Executive for Technology and Enterprise Operations, Patrick Wright adds that experimentation with software development tooling is well underway and already seeing positive results. 

 “I think it’s a really exciting time to be alive. The pace of change is faster than anything I’ve seen in my career,” Wright says. 

“These trials show how our teams can move exponentially faster – getting from requirements, through engineering, development and testing to deployment in weeks or even days instead of months. 

“This puts us in much greater control of our own destiny. Shaping what we want to build and roll out for customers is very quickly going to become much easier.” 

Nicholas Di Giuseppe, NAB Executive for Transformation Engineering leading the Cursor trials agrees. 

“We’ve seen staggering improvements so far. A 40x leap in generating software requirements and a 5 - 6x uplift in development phases. What it means is our teams now have time to ask – ‘what features do we want users to have next?’ – that’s game-changing for them and ultimately for our customers.” 

“NAB is about the customer and that’s where we’re going with this. We know generic offers are not enough - to win you need clever and insightful staff who can curate.” 

Already, creative and practical applications are proliferating. 

Cursor enabled one engineering team to add push notifications to the NAB Trade app 5x faster than expected. Others prompted it to create a referral app for colleagues to help family and friends with banking needs.  The team built a demo-ready prototype compatible with NAB’s ecosystem, design and security requirements in less than 10 minutes. 

Another trial team working to improve merchant onboarding used Cursor to undertake what would usually be six months of progressive elaboration – where teams steadily refine ideas into working software – to successfully develop a host of new features in just 2.5 days. 

A diversified approach 

With the technology advancing rapidly, NAB is taking a diversified approach. The bank was also an early adopter of Amazon’s Q Developer, which began with a pilot involving 350 engineers. 

In one standout pilot case, a team in Personal Banking used Q Developer to upgrade a Java 8 application to Java 17 in just 30 minutes – previously a week-long task. 

Demonstrating the tools’ everyday value, engineers have also used Q Developer to rapidly ‘refactor’ large blocks of legacy Java code – improving its performance, readability and maintainability - simply by describing the functionality they wanted to improve.