NAB creates first AI Science team, appoints George Mathews
NAB has created its first AI Science team and appointed George Mathews to lead it. The unit sits within the bank's Digital, Data & AI division.
The group has been established as NAB expands its use of artificial intelligence across the organisation. It will focus on AI agent architecture, evaluation methods and new AI products, with an emphasis on operating those systems safely in a regulated banking environment.
Leadership move
Mathews joins from QuantumBlack, McKinsey's AI arm. Before that, he worked at CSIRO's Data61 and spent three and a half years at BAE Systems in the UK.
He will take up the newly created role of Executive AI Science, becoming the first person to hold the position at the bank. Mathews holds a PhD in AI and Robotics and a bachelor of aerospace engineering.
The team reports to Chief AI Officer Dr Mahya Knox and sits within the division led by group executive Pete Steel. Its creation adds a specialist function to the bank's technology and data operations as major financial institutions push to turn recent advances in generative AI into internal tools and customer-facing services.
Its work will include technical delivery, with staff expected to write code and build systems and products. Roles in the broader multidisciplinary squad include AI scientists, AI engineers, data scientists, engineers, product managers and designers.
New roles
NAB describes the unit as an example of how AI is reshaping workforces by creating roles that were rare or absent only a few years ago. In the bank's outline of the function, an AI scientist's responsibilities include designing large language model-based systems, agentic AI, natural language processing and optimisation.
The role also covers designing agent reasoning architecture, which governs how an AI agent plans and decides, as well as work grounded in statistics, probability and experimental design. Prompt and context engineering also fall within the remit.
Banks in Australia and overseas have been exploring generative AI for customer service, coding support, document handling and internal productivity. For regulated institutions, the challenge has been balancing speed of adoption with controls around risk, accuracy, governance and accountability.
That has led many large organisations to build more specialised in-house teams rather than rely solely on external vendors or general technology functions. NAB's move suggests banks are defining AI work as a distinct scientific and engineering discipline inside the business, rather than treating it simply as an extension of existing analytics teams.
Industry shift
Knox said the team would support that shift as NAB broadens its use of AI.
"The pace of change being driven by AI is rapid. We see a big opportunity to utilise AI to help us deliver better experiences for our customers, faster," said Dr Mahya Knox, Chief AI Officer, NAB.
"As we implement AI safely and at scale, we must have the right skills, capabilities and experience. We want to build that foundational capability in-house, helping us create safe and valuable solutions that deliver for the whole enterprise," said Knox.
"The AI Science team will play an integral part in building rigorous AI agent architecture, evaluation methodology and new AI products, ensuring the ability to operate these components safely for the whole bank and create outstanding customer experiences," added Knox.
"I'm very pleased to have someone of George's calibre joining us to lead the AI Science team at a critical time for the bank. George has spent a decade at McKinsey's AI arm QuantumBlack and before that had stints with CSIRO's Data61 as well as BAE Systems in the UK," added Knox.
The appointment comes as competition for AI specialists intensifies across banking, consulting and the wider technology sector. Employers are seeking staff with a mix of research training, software development skills and experience in turning models into production systems that can be monitored and governed over time.
For banks, the emergence of dedicated AI Science teams also points to a broader organisational shift. Instead of confining AI work to experimentation or small pilots, institutions are starting to formalise structures around model design, testing and deployment as part of mainstream operations.
At NAB, that means a team charged not only with developing new AI systems but also with defining how those systems should be evaluated before they are used at scale. In practice, this places scientific testing and engineering controls closer to the centre of the bank's AI strategy.
Mathews' background spans consulting, public sector research and defence, giving NAB a leader with experience across commercial applications and technical research.