From cloud stall to cloud success: Bendigo Bank CIO on the power of generative AI

Andrew Cresp, Bendigo Bank’s CIO, shares how generative AI broke their cloud migration stall to deliver genuine industry change.

For the second time in his career, Andrew Cresp, CIO of Bendigo and Adelaide Bank has experienced what he describes as “the cloud stall.”

This is where the easy apps with a strong business case such as the bank’s consumer lending platform have been shifted the cloud.

Other apps that manage payments and branch debtors, for instance, are more difficult to prove a return on investment.

The above video is only an excerpt. Only ADAPT Advantage clients can watch the full video on a Day in the Life of a CIO.

Andrew and his tech team have been able to overcome this issue by using generative AI and the MongoDB open-source database to rewrite application code so the bank can more easily shift some of its older apps to API-enabled cloud infrastructure.

“Fundamentally, what we are seeing is a 90% reduction [in human effort] in the migration of cloud and API-enabling assets. At the same time, it reduces risk [associated with] upgrading the code, but also documenting and creating regression test scripts,” he says.

“So, three things: the modernisation and support, documentation of older apps, and creating a solid regression test script have been the outstanding [benefits] for us.”

Andrew tells ADAPT that generative AI is a “genuine industry change” that is delivering results on the cloud modernisation front – like the rewriting of the bank’s payments and teller system – that he has never seen before.

“We found that gen AI loves the more complex. We found the harder the problem we threw at it, the better it performed. It doesn’t make spelling errors on variables [like humans do].”

Bendigo and Adelaide Bank has been modernising its tech platforms for a few years.

Certain goals have been hit: 50% of its platforms are run in the cloud; the number of core apps has been cut from eight to four; and APIs are being reused across its consumer and business operations.

In May, the bank told investors that it will have consolidated its systems onto one core platform by the end of 2025.

“The ambition realty grew from November to February when we were working through this process, we had a software upgrade of those critical platforms and we said: ‘Let’s go all in.’ We want to [use] APIs for everything and this [will enable the bank to do it] at way less cost than we were planning previously and it’s going to unlock the organisation,” he says.

Building business cases for AI

Almost half (48%) of organisations surveyed this year by ADAPT have indicated that they don’t have any clear use cases for AI.

Andrew advises that organisations need to avoid falling to the trap of using AI solutions to look for a problem, instead of the other way around.

“Our business problem was we were getting to about 50% [through the] cloud migration and we were going to struggle to find a way to finish the job off.”

Using gen AI to help with this its cloud migration was a compelling use case because the bank doesn’t have to use customer data or shift data outside its walls.

“That was strong for us,” he says.

Meanwhile, Andrew says he has asked the bank’s service desk team to identify three problems that can be solved by using gen AI.

These would centre around how to improve processes for internal staff and external customers.

The bank services 2.5 million retail customers through a network of around 500 branches and agencies.