Incident sheet · /infrastructure

Filed
Read
3 min · 593 words
Severity
Routine

ANZ moved 1.7PB to AWS in six months with AI assistance

ANZ moved 1.7PB, 300 million files and 4,500 workflows to Amazon EMR in six months, using agentic AI to cut migration effort by 35% and halve defects and rework.

ANZ uses AI to move 1.7PB data platform to AWS
FeedANZ uses AI to move 1.7PB data platform to AWS — RodDIYguy / pixabay

Event log

  1. ANZ moved 1.7PB to Amazon EMR in six months, versus the 18 months it had planned.
  2. The estate included 300 million files, 142,000 tables, 4,500 workflows and 4,000 views.
  3. Agentic AI cut migration effort by 35% and defects and rework by about half.
  4. Almost 80TB was tokenised, cutting operating costs from over US$10,000 a day to under US$3,000 a month.
  5. ANZ used internally approved but unnamed models in its Amazon Bedrock troubleshooting assistant.

By Olivia Hart

ANZ Banking Group moved 1.7 petabytes of data to Amazon Web Services in six months, a year faster than the 18-month schedule it had planned.

Inside the project’s “hothouse environment,” chief data and AI officer Kai Yang said at an AWS media briefing that business and technology staff sat in one room with AWS engineers. The team used coding agents and built troubleshooting agents to reduce defects.

The bank took one of its largest enterprise big-data platforms off an on-premises Cloudera Hadoop environment and onto Amazon EMR, AWS’s managed service for open-source data-processing frameworks such as Hadoop and Spark.

Previously called Elastic MapReduce, EMR runs those frameworks on rented clusters of virtual servers, sparing customers the work of building and maintaining the hardware themselves.

The estate comprised:

  • 1.7 petabytes of data.
  • 300 million files.
  • 142,000 tables.
  • 4,500 automated workflows.
  • 4,000 views.

ANZ largely used a lift-and-shift approach, moving workloads to the cloud with minimal redesign. When iTnews asked what its comparisons were measured against, the bank cited the estate’s scale and complexity, including legacy technologies and complex access controls.

ANZ estimated that agentic AI reduced migration effort by 35% compared with traditional approaches and cut defects and rework by about half. Engineers also used a troubleshooting assistant built on Amazon Bedrock to find and fix performance bottlenecks.

Group chief information officer Donald Patra said the project showed AI-assisted engineering could “accelerate complex delivery while maintaining a disciplined focus on security, resilience and quality.”

ANZ did not name the models in the Bedrock assistant, saying only that the bank had internally approved them.

The platform now supports the bank’s payments, customer remediation, and business and private banking operations.

Before the migration, ANZ had to protect sensitive customer data. It used tokenisation, replacing sensitive values with stand-in tokens, and has tokenised almost 80 terabytes so far.

The bank said the capability ran about seven times faster than its initial approach. Tokenisation operating costs fell from more than US$10,000 a day to less than US$3,000 a month. ANZ confirmed that those figures cover tokenisation alone, not the platform as a whole, and did not describe the earlier approach.

AWS Professional Services and unnamed AWS partners worked on the migration alongside ANZ’s own engineers. Chief executive Nuno Matos said last year that the bank would depend less on third parties, using them only in special cases such as speeding up delivery.

Yang said the modernisation was necessary to give ANZ good-quality data and a scalable lakehouse architecture, combining data-lake storage with warehouse-style management. He also said the bank is building a reusable AI platform designed to run on premises as well as on cloud services such as AWS.

At its annual results last week, ANZ revealed an enterprise AI control plane for its AI systems. Yang said it would provide continuous monitoring, auditability and human accountability.

What does the migration change for ANZ?

For ANZ, the change is practical: core workloads now sit on a managed cloud service, while the project supplies a measurable AI-engineering benchmark—35% less migration effort and roughly half the defects and rework.

The wider change is governance: reusable AI tooling will run across on-premises and cloud environments under continuous monitoring, auditability and human accountability.

For technology leaders, the result is a clear benchmark: a year saved, 35% less effort and half as much rework, with security, resilience and quality kept explicit.

via itnews.com.au (Source)

Filed under

  • anz
  • amazon-web-services
  • cloud-migration
  • big-data
  • agentic-ai
Share this article:

More from Olivia Hart

Olivia Hart

Show full bio

News editor covering money, funding and market shifts around enterprise information technology and the leaders who run at CIO Monitor.

207 articles

The newsroom digest

One email a day with the stories that matter. No noise.

By subscribing you agree to the privacy policy.

Cross-references

‹ Previous articleNext article ›