Published on the 12/09/2025 | Written by Heather Wright

What hit the spot for A/NZ leaders…

Helius Guimaraes, Fonterra chief data and AI officer, sums up Gartner IT Symposium/Xpo in a succinct sentence: AI is everywhere.

Speaking at the conference on the Gold Coast this week, Guimaraes says his key takeaway from the conference is the investment required for change management if you’re doing an AI project.

Alicia Mullery, Gartner vice president, analyst, warned of a ‘transition mortgage’ in AI.

“There is a real limit on what you can achieve with cost reduction. But value creation is much more unbounded.

She noted that CFOs are losing track of what is being spent on AI projects, with 65 percent of Australian and New Zealand organisations barely breaking even, or losing money on their AI investments.

“When you first implemented ERP you licensed the software and trained your people on one kind of work. But with AI the work and the training keeps switching context,” Mullery said.

Workers are likely to use AI in multiple scenarios from summarising emails, to doing data analysis and creating videos, using different models along the way.

“Most people are just discovering what they can do. They’re not being trained on what they should do.”

Darryl Plummer, Gartner VP, distinguished analyst noted the initial AI implementation costs were just the start, saying companies had to invest heavily in AI training and literacy programs and change management.

For every 100 days of implementation AI adds 25 more to train staff, he says.

“And worse yet, change management costs another 100-200 days – that’s an up to 200 percent more effort.”

Ancillary costs will also add up, he warned, noting that for every one AI tool purchased, Gartner says there will be 10 hidden costs companies didn’t anticipate, such as managing access credentials for agents, acquiring new data sets to ground AI, managing multiple models or packing your ‘accuracy survival kit’.

“For every AI tool you buy, anticipate 10 hidden costs. Conduct an analysis and decide which costs you will fund. The last thing you want to be is the unintended owner of a negative ROI business case,” Mullery says.

Guimaraes says that stark warning will change the way he approaches AI at Fonterra.

“We are going to need to basically invest more on change management,” he told iStart.

“Just putting in an AI solution and thinking everyone will adopt is not enough. You need to take people on the journey by understanding, from management to the users, how AI is going to benefit their job.”

Dan Cooper, business design and architecture lead for insurer Suncorp Group was also eyeing the cost equations, noting a big concern from corporates he had spoken to that AI was all about cost reduction.

“I really like that a lot of the messaging from Gartner is that it’s actually about value creation. There is a real limit on what you can achieve with cost reduction. But value creation is much more unbounded. You just have to have a little bit of bravery.”

Hung LeHong, distinguished VP analyst says it’s very hard – though not impossible – to get financial value from AI. Just five percent of CFOs say they saved money from an AI initiative, while six percent said it increased revenue or profit, and 74 percent said it saved time.

“Time saved is not money saved.

“Productivity is a low-quality benefit. It needs much effort to turn it into value.”

James Anderson, Gartner VP analyst, urged attendees to map IT measures to business priorities, using metrics to measure IT’s impact on what matters to executives, such as increasing customer satisfaction and retention, risk reduction, becoming a market leader, growing revenue or reducing cost through process efficiency.

Welcome to the wild west

The wide array of uses for AI, highlighted in sessions, on the conference expo floor where vendors clamoured to showcase their latest and greatest AI wares and in chatter among attendees, highlighted a ‘big disconnect’ between what AI is, what it really means and where it will be used, Cooper says.

“There are a lot of people trying a lot of different things and it is going to be a while before we figure out which of those things stick and which were a great idea but either people or technology are not ready for it,” he says.

“And I don’t think there is a way to tell right now. It’s a whole wild west of what people are doing with this new technology.”

Cooper, who studied AI in 2001 and was told back then ‘we were *this* close to a break away’ noted the marketing hype – “I’ve literally got the t-shirts,” he joked (and yes, there were plenty of AI focused t-shirts being offered around at the event).

“Time will tell the use cases that stick and we have got to discover them. There is no recipe for success… you have to just put yourself out there, try some stuff, see what lasts, what works.”

That was something being mulled over by Daniel Camilleri, head of IT at Yamaha Motor Australia, who said a key takeaway from the conference for him was how to incorporate a better use of AI within the organisation.

“We need to understand what benefit we need from AI. Do we want to increase productivity? Do we want to increase accuracy of what we do or improve head count?

“Once we determine that we can try different ways of incorporating AI whether it is automation for productivity or using data to improve accuracy,” he says.

“We need to determine the value of AI and if that takes precedence over other activities.”

Camilleri says he’s not feeling pressure from above to implement AI asap, saying instead the focus is instead on getting business outcomes first and foremost, whether through AI or another technologies.

“As some of those pressures are made more complex and more pressing, that’s where we see AI potentially playing a bigger role.”

Not everyone is ready

He’s among the lucky ones not feeling pressure to implement AI asap, and at any cost.

The head of technology for one government trading enterprise, who spoke on condition of anonymity, said hype was driving panic among the board of her organisation, which thought implementing AI would solve all its problems.

“Our board thinks we will go out of business if we don’t adopt AI. But we are a government agency, so we will not go out of business, and we haven’t even automated or standardised business processes. They are panicking, but we haven’t even go the normal basics right.”

Gartner’s Mullery noted the odds of an AI initiative achieving ROI were one in five, with the odds of AI achieving  true transformation being one in 50. She also cited the recent MIT report which claimed 95 percent of genAI projects and pilots are failing.

Says the government IT leader: “If you rush something just because you don’t want to be left behind and the foundation isn’t right, you may as well just open the window and throw the money out. Then at least the person in the street would benefit!”

She said she would be using much of the content from the event, particularly around what needed to be considered when implementing emerging technologies and what concrete aspects do you need to have in place to do AI, to help educate her board so they didn’t get carried away too much with ideas of not existing if they didn’t do AI immediately.

“I’ve been trying to tell them about AI literacy, but I didn’t have enough material. Now I feel I have all of that,” she says.

New legislation coming this year will require her organisation, as a government organisation, to keep a log of all AI done to ensure transparency and the ability to show clients where in the process the technology was used.

“Our users don’t understand yet that you can’t just trust the output of AIs. They don’t understand the concept of probability.

“We are in the era of AI, and it is here. But that doesn’t mean we as an organisation will be able to adopt.”

A colleague from the same agency noted too that the organisation had a ‘quite unique’ business model and lack of standardisation which would further hinder AI adoption.

Bihma Jayasinghe, chief operating officer of Melbourne-based healthtech Data Capture Experts says one of her key takeaways was the issue of aligning leaders with the changes coming, when those changes are still a little grey and uncertain.

“There is a lot of education that needs to happen still.”

As a healthtech provider, she says for her business the conference strengthened her belief in the need to help educate the healthcare sector.

“We need to be working with the CIOs in the sector. They have a big role in this space and carry a massive responsibility with the decisions they are taking on behalf of the organisation, or what they are taking back to boards, who are not necessarily technical enough to understand the whole context, for approving.

“From a healthtech point of view there is a lot of data out there which needs to be rich for those AI efficiencies to come in.”

Jayasinghe also called for all leaders to think about the human impacts of the technology.

“We still need as leaders to be thinking about the families, people working for you and how we help them if their positions are to be challenged or there is reutilisation of their time and resources. It’s an important factor to think about.

“We get inspired by technology and the new things that can be done. But we shouldn’t forget who we are though.”