From AI experimentation to enterprise capability

As organisations move beyond AI experimentation, there are different approaches to embedding capability into their workforce. Some have started small before widespread implementation; others have opted for a short, complete immersion; and all have valuable lessons to share.

During the recent CommBank 2026 Accelerate AI event, CommBank Chief AI Officer Ranil Boteju noted that rather than a pure technology issue, AI has evolved into an immediate leadership priority, centred on execution, capability and trust.

“Increasingly, the question is no longer whether AI matters, but how organisations apply it responsibly, at scale, in ways that create genuine value for customers, communities and the broader society,” he says.

“Increasingly, the question is no longer whether AI matters, but how organisations apply it responsibly, at scale, in ways that create genuine value for customers, communities and the broader society.”
– Ranil Boteju, Chief AI Officer, CommBank

A phased approach

Retail giant Coles was an early adopter of AI – with a dedicated machine learning and AI team for close on 10 years. Coles CEO Leah Weckert explains the company now has significant AI use cases, including within its replenishment system.

“Every product that you buy when you come into Coles has made its way to the shelf via machine learning and an AI system that drives that,” she says.

“Every product that you buy when you come into Coles has made its way to the shelf via machine learning and an AI system that drives that.”
– Leah Weckert, CEO, Coles

“We have about 26,000 different products across our 850 stores and the machine learning system is forecasting where those products need to be and in what quantities 100 days out, so we can place supply orders.”

Weckert also highlights the importance of team training when introducing AI tools. Last calendar year, Coles took a phased approach to roll out ChatGPT and Copilot for store support-centre team members.

“We invested a lot in a phased rollout, where we took about 500 team members to start and we put them through intensive training,” she explains. “We made them champions in their functions and we gave them support to find early use cases, so that when we went live with everybody, that group of people were advocates and champions around the business.”

Going all in

While some organisations start small with AI, others embark on widescale transformations.

Ben Chan, Chief AI Officer at data and analytics company Quantium, says the company recently underwent a “massive cultural transformation” to go all in on AI. He says the program was implemented “over a couple of months, through a lot of hard work and a lot of top-down leadership pushing AI through the whole company”.

The result? “92% of all Quantium team members use AI every single day,” he says.

Quantium’s transformation strategy included a process called “AI shares”, which helped to normalise and embed AI usage.

Explains Chan: “At the start of every major meeting, we would have an ‘AI share’, where people would share something they’d produced with AI or even something that didn't go so well with AI.”

He also emphasises that access to AI does not guarantee competitive advantage.

“What does give you the advantage is when you start connecting your IP, your data and how you do things as an organisation, and I think that really starts to build a sustainable competitive advantage,” he says.

“What does give you the advantage is when you start connecting your IP, your data and how you do things as an organisation, and I think that really starts to build a sustainable competitive advantage.”
– Ben Chan, Chief AI Officer, Quantium

Chan provides a hypothetical example of a retail client connecting customer transactions and inventory through AI.

“We're going to ask, ‘Hey, Claude, how has coconut water performed at the Vermont South store over the last 13 weeks versus last year?’. AI goes into the systems and into the metrics layer that we’ve built, and comes up with the numbers for that store,” says Chan.

“Think about how much time that saved you and your team. You don't need to wait until the next day for your team to get back to you with a dashboard. It also overlays AI insights around what you can do about your sales.”

The data challenge

Research from IBM shows AI is only as good as the data it is built on. As Chan puts it, “it’s rubbish in and rubbish out”.

“[Data] needs to be a single source of truth and AI has helped that a lot – no longer do you need the big, three-year data migration and data-cleaning exercise because AI has sped that up a lot,” he says.

Sam Hemphill, Executive General Manager, Customer, Channels and Data, CommBank, notes the challenge that data presents to AI implementation. He says companies can often pause an AI rollout until their data is in ideal shape, or proceed while hoping AI helps clean the data during implementation.

“We certainly feel you must do both,” he says. “One without the other is not going to be sustainable in any way. I think the intersection of the two is where businesses really need to be focused.”

AI presents huge opportunities for organisations, but Boteju stresses that the value depends on leadership choices about where to deploy AI, and how to build capability responsibly and at scale.

“When organisations demonstrate that AI can be governed effectively and deliver visible benefits, adoption will accelerate until it becomes trusted, embedded, invisible infrastructure for the economy.”

“When organisations demonstrate that AI can be governed effectively and deliver visible benefits, adoption will accelerate until it becomes trusted, embedded, invisible infrastructure for the economy.”
– Ranil Boteju, Chief AI Officer, CommBank