Meet Adapt, Westpac's Azure-based enterprise data platform
Key points
- Westpac has unveiled Adapt, a new enterprise data platform on Microsoft Azure that brings together data from 285 source systems into one common governed environment.
- The Westpac Intelligence Layer, built with Snowflake, creates a single customer view across products and channels by running AI models on data at rest without moving it.
- Westpac’s broader technology ecosystem includes Microsoft, AWS, Snowflake, Nvidia, Google and AMP Frontier, with partners supplying forward-deployed engineers and Nvidia powering an in-house AI factory.
Westpac has revealed a new enterprise data platform that runs on Azure and acts as a foundation for its customer intelligence activity and AI work.
Its two key executives – chief data, digital and AI officer Andrew McMullan and chief AI officer Dan Jermyn – on Tuesday offered the most complete view of the bank’s data and AI setup.
A lot has been said about this extended setup over the past year, but until now it’s been hard to conceptualise what all the different layers and tools do.
At the heart of Westpac’s setup is a platform that it has previously hinted as having, but never expressly talked about publicly: a new enterprise data platform (EDP) that it is calling ‘Adapt’.
Adapt, McMullan said, represents “one of the largest data modernisation programs ever undertaken at Westpac”.
“AI only scales with trusted data. That’s why we’ve spent the last few years simplifying and standardising the way data is managed across the group,” he said.
“Adapt is our new enterprise data platform. It brings together data from 285 source systems into one common governed environment on Microsoft Azure.
“Delivering it required the migration of over 1PB of data and more than 14,000 data pipelines onto one single platform.
“Having this one enterprise data platform in the Azure cloud helps reduce duplication, allows us to apply governance much more consistently, and gives our teams faster, more secure access to the data that they need.”
Data source migration into Adapt was completed in March this year.
“It was ‘hard yards’ to get the platform to where it is, and the quality and reliability of the data there is a real step change for Westpac,” McMullan said.
“[The] 285 systems is the majority of the data we need to be able to understand and serve our customers at Westpac better.
He added that iInternal “teams who use [Adapt] are moving faster than they’ve ever moved before.”
The creation of Adapt is closely linked to Westpac’s broader Unite program, which is effectively re-platforming the various parts of the group onto a simplified technology stack.
“Adapt and Unite work together,” McMullan said.
“Unite is simplifying the underlying systems, [while] Adapt connects and governs the data across them.
“Together we’re creating a simpler and stronger foundation for how information flows through the bank, and that’s really important because every digital experience, every insight and every AI capability depends on the quality of this foundation.”
Customer intelligence
One of the platforms that relies on the existence of Adapt is the Westpac Intelligence Layer, which first broke cover last year.
It’s been unclear for some time what role the Westpac Intelligence Layer plays in the data and AI ecosystem.
McMullan clarified this: in simple terms, it’s a customer intelligence platform, creating a single view of customers, the different financial products they use, and the “signals” that their different interactions with the bank generate.
“One of the most important opportunities created by Westpac Intelligence is the ability to bring the whole bank to the customer,” McMullan said.
“Historically, customer relationships have been fragmented across products, channels or business units. A customer may have a mortgage, transaction account, a business relationship, maybe an investment relationship with us. These interactions have often been managed separately.
“Westpac intelligence connects the full customer, not just individual interactions, giving us the ability to understand, anticipate and serve customer needs across every single touchpoint.
“It brings together customer relationships, behaviours, their products, transactions and interactions into a more complete view of our customer.
“The goal is understanding the full customer relationship to enable a more relevant connected and timely response across all our channels.”
Snowflake plays a significant role in the Westpac Intelligence Layer.
“In partnership with Snowflake, we’re underpinning our trusted data foundation [Adapt] with our intelligence capabilities,” McMullan said.
An important architectural feature of the Westpac Intelligence Layer, according to McMullan, is that “it allows us to make intelligent decisions without needing to move data.”
“We can now run the most advanced machine learning and AI models across our data at rest, meaning we’re now in the business of using our data to better serve our customers and not moving data [to feed AI models or use cases],” he said.
McMullan added that this architecture is helpful for incorporating data that, for specific reasons, is stored in other hyperscaler ecosystems.
“Our main data platform is on Azure but we also do things with Google Cloud and with AWS, and the reason that we chose Snowflake as the intelligence layer on top of that is because we can be intelligent on the data at rest without moving it across those hyperscalers,” McMullan said.
Broader technology ecosystem
Westpac revealed details about how it works with an ecosystem of technology partners in the data, digital and AI space.
These include Microsoft, AWS, Snowflake, Nvidia, Google and AMP Frontier.
One of the key things that Westpac is looking for from its partnerships is access to engineering talent.
Increasingly this comes in the form of “forward-deployed engineers” that hyperscalers are employing to help end users embed particularly emerging technology faster, with the aim of realising tangible value sooner and justifying investment and spend.
Microsoft has nine forward-deployed engineers “working with over 100 [Westpac] data professionals”; AWS is also supplying an unspecified number of these engineers as well “to automate our backend operations” with respect to the infrastructure required to operate and deploy AI agents.
Nvidia AI factory
A technology partnership Westpac has not spoken about much is its work with Nvidia, which powers a so-called “AI factory” for the bank.
“Nvidia technology powers our AI factory, allowing us to run advanced AI workloads much closer to our data with security, control and performance as required,” McMullan said.
“Their leading AI engineers work side by side with our team.
“Together, we’re creating the foundation to develop, train and deploy AI at scale while maintaining the governance, resilience and trust our customers expect from us.”
About Author
