Exclusive: UiPath sees ROI pressure in ANZ AI shift

Australian and New Zealand enterprises are accelerating their shift from exploratory AI pilots to production-scale agentic systems as boards demand clearer returns and tighter governance, according to UiPath's local lead.

"Executives, business executives, need to see real ROI from their investments," said Peter Graves, Area Vice President, ANZ, UiPath.

"They've been happy to put some money into pilots and experiments. But now they're asking, where's the ROI? How do we productionise this? Where's the real value?"

Graves said organisations are consolidating tools and becoming more selective. "A lot of them are looking at consolidating tools and platforms, not adding more. We've got one of everything. It's not like we don't have enough technology. We've got too much," added Graves.

He said large enterprises are rethinking their architectural capability models and assessing which platforms can support long-term transformation.

"They definitely need to make change, but where to place their bets and which vendors and technologies to place bets on is challenging," added Graves.

Layered AI

Graves said organisations are adopting more structured approaches to deploying AI. "You can't go from zero to 100 and expect employees to get it straight away," he said. "There's a lot of fear, and that can lead to low adoption and then low return on investment."

From a technology perspective, he said businesses are combining foundational models, internal data investments and vendor capabilities. "What UiPath is trying to do, we're trying to be very agnostic, and we're trying to provide an agile orchestration layer that can sit across an end-to-end business process," added Graves.

He said the most advanced organisations are no longer tweaking existing processes.

"The best companies are reimagining the business processes from the customer or the employee back with agentic in mind," added Graves.

"The leading adopters of agentic have moved from project based task optimisation to program based process optimisation, orchestrating existing technology investments along with agents, robots and humans - this is where measurable earnings before interest, taxes, and amortisation (EBITA) contributions are being made by CIOs."

Australian adoption also remains high.

"Australia was definitely the leading region, and we were about 90% of organisations have adopted or are planning to adopt some form of agentic AI," said Graves. He attributed this to the region's mature technology governance.

"Australian organisations are just able to take advantage of it a little quicker than others, and that's why you're seeing those numbers," he added.

Governance first

Graves said the gap between strong performers and stalled initiatives comes down to foundations. "The organisations that are getting the most value and seeing the best results have spent the time to set up the right security, trust and governance framework," he said.

Initiatives can falter when these structures are absent.

"It can't get past that pilot stage because it hasn't ticked the boxes of whether it's security or they need access to a different data set, or they haven't got an AI policy," added Graves.

He said most leaders are now pragmatic. "They've seen the challenges of taking something from a pilot or a POC into true production without those guardrails. It's almost impossible," added Graves.

For agentic systems, the stakes are higher. "Unless you give it access to the tools, it won't be able to achieve the goal," said Graves. "That is the number one thing that is stopping programmes from scaling."

Industry leaders

Financial services, insurance and healthcare are among the most advanced sectors.

"A lot of the large insurance providers are already automating the claims process because it is highly manual with lots of documentation," said Graves.

Healthcare is also accelerating. "Gold Coast Health has invested heavily in automation, and now they're adding an AI and agentic layer to take on that last mile that traditional automation couldn't do," added Graves.

He said public sector momentum has grown sharply.

"The opportunity to drive real citizen outcomes through the use of agentic is huge. Collapsing wait times for visas or passports or any kind of government service is huge," added Graves.

UiPath often references Woodside Energy as an example. "They've deployed our communications mining tool and collapsed the time to deal with incidents dramatically," said Graves. Many organisations still rely on manual email triage. "It's very surprising how often you find that in really large organisations," added Graves.

Scaling strategy

Graves said the most successful customers treat agentic AI as an organisational capability rather than a sequence of standalone projects. "The customers that are at the leading edge are actually building this as an internal capability," he said.

Workshops and cross-functional planning are critical.

"You tend to walk out of those meetings with 10, 20, 30 different use cases. And then it's about prioritisation," said Graves. Quick wins often underpin funding. "We tend to try and create self-funding programmes," added Graves.

He said more enterprises are shifting from task-based automation to broader roadmaps. "They're building a team, they're building a capability, and they're saying, here's the 10 things we're going to do over the next 12 months," added Graves.

A leadership mindset

For executives planning ahead, Graves said adaptability is essential. "You need to adopt an agile mindset," he said. "You need to build a very agile company that lets you adapt, try new things, learn quickly, and then when you find something that works, double down and move very fast."

Graves said moving too quickly without preparation "carries risks."

"Unless you really manage that properly, there's a real risk it can backfire, and the returns you're seeking won't be achieved," added Graves.