50 AI squads target growth and productivity…
National Australia Bank is rolling out an Anthropic-powered AI platform to more than 40,000 employees and has established 50 business-led AI squads as it seeks to embed the technology into lending, deposits and customer-facing operations.
During the bank’s FY26 full-year results briefing, NAB group chief executive Andrew Irvine described AI as ‘one of the most consequential technology revolutions that we’ve had in humanity’, as consequential as the steam engine, electricity and the computer.
“As a bank and as a highly regulated entity, we have to get it 100 percent right with AI.”
The move highlights a shift from experimentation to a coordinated operating model combining an enterprise-wide single AI and agentic platform with 50 business-owned implementation squads focused on use cases ‘where we see significant customer, colleague or shareholder value to prosecute’ and aligned with the bank’s three business priorities of growing business banking, deposits and proprietary home lending. Programs with AWS and Anthropic have been used to upskill the AI squads.
“We’re leaning in hard and deploying capability there,” Irvine says.
He noted the example of using the internal platform, called Ruby, to enable business bankers to ‘very, very significantly’ speed up the process of writing credit memos. “The bankers are still very highly accountable for that credit application, but they can leverage the power of Ruby and AI tooling that will be offered to them.
Ruby Monday
More than 40,000 staff will have access to the Ruby platform. The announcement comes hard on the heels of the bank announcing it plans to cut 2,500 staff – or six percent of its workforce – by June 2027 as part of a major restructure. Around 1,500 will lose their jobs by the end of this year. No mention was made of AI in that announcement. However, NAB said the moves would help make it a stronger bank, better positioned to navigate the economic conditions, customer expectations, competition and rapid technological change. The Finance Sector Union, meanwhile has raised concerns about the ‘lack of transparency around AI’s role’ in the decision.
In its investor presentation, NAB, which owns Bank of New Zealand, said it has a ‘multi-year ambition to modernise our technology to build a bank that is simple, fast and resilient,’ positioning technology modernisation as a core enabler of future growth and operational performance.
Irvine says the bank expects augmentation, rather than full automation, to deliver the most immediate value when it comes to AI.
“As a bank and as a highly regulated entity, we have to get it 100 percent right with AI. And so, for now, having augmentation use cases where humans are still completely accountable is going to be very important before I’m comfortable with more automation use cases,” he says. He acknowledged, however, that the automation use cases are expected to have a bigger impact on the workforce.
NAB says agentic AI’s benefits flow in two key areas for the bank: Growth – in enabling better conversations, faster decisions and higher value outcomes – and productivity.
On the growth front, it says early value outcomes include a 20 percent reduction in the time to prepare business lending credit memorandums, and a 75 percent reduction in the time to identify customer insights, enabling bankers to bring ideas to customers sooner.
NAB says early value outcomes also include a 50 percent reduction in customer lead preparation time for personal bankers, while on the productivity side they include a 50 percent reduction in design to code delivery with greater accuracy, 80 percent faster triage of cyber incidents and a 90 percent reduction in technology cyber assessment time, accelerating issue detection.
The next wave of agentic tools, including voice-enabled servicing agents, complaints case management automation, enhanced fraud detection and resolution and ‘financial crime detection agents’ is in active build.
Foundations first
The bank has been setting the foundations in place in a multi-year transformation project, which has included decommissioning around 600 apps since FY24 and transferring around 90 percent of apps to the cloud. A modern data platform, single sales platform and single customer master which supports 96 percent of the bank’s personal and business and private banking customers are also part of those foundations.
This financial year also saw the creation of the company’s first AI science team, headed by George Matthews – ex McKinsey Quantum Black, CSIRO Data61 and BAE Systems, as inaugural executive AI Science and the launch of a development toolkit and framework for agentic AI.
Databricks Genie was also rolled out, providing AI enhanced analytics for staff in a move the bank says saw savings of two to four days of development time per use case in early use, with further gains expected as adoption grows.
By combining AI with advanced analytics, NAB says it can analyse information from customer interactions, transactions and other banking data to build a more complete picture of customer needs. The capability underpins the bank’s ‘Customer Brain’ platform, launched in 2023 and designed to deliver more relevant and timely communications.
Jessica Cuthbertson, NAB executive for data and analytics, says integrating multiple data sources can improve customer support during critical moments. She cites the example of transaction disputes, saying it enables the bank to keep customers proactively informed at each stage and noting that customers who are proactively informed throughout the process are 30 percent less likely to call to get information or peace of mind.
