NAB and Microsoft design AI ATM

National Australia Bank (NAB) and Microsoft have designed a proof of concept automatic teller machine (ATM) using artificial intelligence and cloud computing technologies.

The cloud-based app, which uses Azure Cognitive Services, removes the need for physical cards or devices to access cash from ATMs. Instead, a customer who opted into the service would be able to withdraw cash from an ATM using facial recognition technology and a PIN.

NAB chief technology and operations officer Patrick Wright said the ATM is a look into what the future might hold for the way customers access banking products and services.

“Cloud technology allows us to take advantage of features and capabilities that are world-leading and enable us to deliver at pace for our customers,” Wright said.

Microsoft’s local boss, Steven Worrall, said cloud computing and AI present the opportunity for a new generation of financial services to be developed and deployed at scale.

“We believe AI will profoundly impact financial services and the sorts of solutions that banks will be able to deliver in the future. For a consumer-facing application such as the AI-powered ATM we’ve developed with NAB, this sort of continuous AI innovation is important,” he said.

In a blog post on the Microsoft website, NAB said one of the first things the bank has focused on in its digital transformation is how to transition to the cloud.

NAB’s executive general manager for infrastructure, cloud and workspace, Steven Day, said the bank is trying to identify patterns and then deploy a set of automated build patterns and landscapes for its developers. This allows the most critical workloads to be safely deployed in a true multi-cloud environment.

“On our less critical workloads, we can safely deploy them to a single provider,” he said.

Yuri Misnik, executive general manager and CIO at NAB, added that every time the bank builds or buys an application, the default approach is public cloud, ideally using native cloud services.

“If that’s not possible – could be regulatory, application workload-specific reasons – we consider alternatives,” he said.

The second pillar of the strategy is to use APIs and streamline operations by taking a consistent approach to building code, deploying code to production and managing applications. The third pillar, Misnik said, involves maximising the use of the bank’s operational data and applying modern tools and machine learning algorithms to it, which enables engineering teams to run in a true DevOps way.