ANZ has shed light on its nearly two year old drive to make its automation efforts more repeatable, helping more of them move beyond experimentation and reach production scale.
The bank first revealed the its decision to “industrialise” its process automation efforts back in April 2017, though it did not respond at the time to questions on how this would play out.
However, ANZ’s group executive of technology, Gerard Florian, told the bank’s own podcast this week that the establishment of the industrialisation domain was progressing.
“What we’ve done [is built] a team - a domain actually - that is focused around business automation and integration, which includes things like digital assistants, basic document digitisation but also APIs and so forth,” Florian said.
“We’re trying to think about how we will build out that [capability as a] factory so that instead of embedding a little bit of automation in lots of separate projects but never getting repeatability and never getting to scale, we can have an environment where different projects can basically go to the ‘automation app’ store, if you like, and pull things out that they can use in their projects, so that we get better as an organisation.
“We’re trying to get more of that repeatability in there.”
Florian raised an example of ANZ’s chatbot and digital assistant efforts.
“We have - if I think about just chatbots - within the bank lots of people experimenting with chatbots,” he said.
“I think we had in excess of 16 little pilot projects happening around the place - which is great [because] you want people to be curious, you want people to be innovating.
“The challenge is they won’t get to scale unless we can consolidate effort and build a certain element of enterprise capability.”
Florian indicated another project worked on by the automation domain combined digitisation and machine learning.
“How do we ingest documents that have to be paper-based - for example trade documents with our customers that have multiple signatories? How do we digitise them, but then use machine learning and OCR technologies [on them] to get better and better at the checks that take place on those documents?,” he said.
“[We’re] still having people involved in the process, but perhaps less than previously.”
The bank has set some limits on automation, however.
CEO Shayne Elliott told the SIBOS conference in Sydney late last year that “no robot can replace a true relationship with a banker” - suggesting, as Florian does, that people will still be involved in partially-automated processes.
Cloud adoption slower than anticipated
Florian used the same interview to provide an assessment of ANZ’s cloud journey to date.
“Our journey at ANZ hasn’t moved as quickly as perhaps we would have liked, to be very open,” he said.
“Some of that can be internal restrictions, some of that can also be external - clearly we’re a highly regulated industry and there’s a role there to be played to make sure that the direction we’re taking around cloud is in line with where the regulators would like us to be.
“The good news is that over the last couple of years, both internally and externally things have continued to evolve.
“So I’d say the journey we are probably more about to embark on versus the one that we have been on is fairly clearly around extending our enterprise into the cloud.”
Florian said that ANZ saw cloud as an extension of capabilities and workloads it would still run in its own data centres.
“We will still have data centres and infrastructure … [but] we would like to be able to extend that into the cloud to make use of PaaS, SaaS - it doesn’t matter - but [we want to] make sure we’re in a position to use far more of the capability that is being provided.”
ANZ’s position was driven by having “a lot of environments that are less likely to be easily modernised”, Florian indicated.
“Let’s look at workloads, let’s make some sensible decisions both commercially and technically, and let’s also start to think about guiding principles for the new applications that we’re delivering,” he said.