AI transformation is '10% technology, 90% people'

Two senior figures driving artificial intelligence (AI) adoption inside their organisations have reached the same conclusion from opposite sides of the boardroom table: the technology is ready, but the workforce is not.

In separate interviews with HRD, Cathy Doyle, chief people and culture officer at Sydney-based life insurer NobleOak, and Rochana Golani, vice president (VP) of learning and enablement at data and AI company Databricks in California, both argued that AI transformation is fundamentally a talent problem dressed up as a technology one.

"It's not a technology problem, it's a talent problem," Golani said. "Every organisation that I talk to keeps bringing this up: 10% technology, 90% people."

The chief people officer as 'utility player'

For Doyle, that shift is reshaping her own job description. As NobleOak's chief people and culture officer, she now oversees the insurer's people function alongside its use of artificial intelligence and AI agents. "I look after all of our people and AI agents – I look after AI as well as technology, as well as people and strategic projects," she said.

She sees that dual mandate becoming the norm rather than the exception. "No longer can you just be one hat to wear in an organisation," Doyle said, describing the kind of multi-skilled executive she calls a "utility player" – someone equally credible on commercial strategy and technical capability. Chief people officers (CPOs) who can't demonstrate both, she argued, risk being sidelined as businesses reorganise around AI.

That doesn't mean every CPO needs to become a data scientist, Doyle was quick to add. "I'm not a cyber specialist, but I know I need a head of cyber," she said. "You've just got to go back to your disciplines, what environment am I working in, and therefore what do I need to cover, and then what skill sets do I need to go get."

Why managers, not executives, decide if transformation sticks

Golani's view from inside a fast-growing technology company points to a similar conclusion, delivered from a different altitude in the organisation chart: leadership commitment matters, but the real transformation happens, or stalls, with frontline managers.

"Managers to me is where this work starts and how we drive that transformation," she said. "I think of managers as the biggest change agents that do this well." She argued that manager expectations need to shift toward active role-modelling of AI use, coaching teams through the change rather than simply mandating it from above.

That philosophy has shaped how Databricks talks to its own customers about workforce change. Rather than hiring in new AI-fluent talent, Golani said the more successful organisations she works with retrain their existing people.

She pointed to National Australia Bank (NAB), which she said is running am AI bootcamp, as an example of upskilling at scale rather than a hiring spree.

Golani also drew a distinction between organisations merely "scaling" old habits with new tools and the smaller group genuinely reinventing how they work. "A very large percentage of people today are what we call 'scaling' – they're doing the things that they used to do, they're just doing it with AI," she said.

Genuine reinvention, that is creating new revenue streams from AI rather than just accelerating existing processes, is rarer: "it's less than 5% of companies that are able to do this," she estimated.

Curiosity over budget

Doyle's own experience underlines how deliberately some organisations are investing in AI literacy at the top. She described taking NobleOak's CEO and a small executive group to Silicon Valley for an intensive AI immersion.

"The CEO and five of us, we went and did an AI course, we went to Silicon Valley to find out what Silicon Valley thought of AI, and went to Google," she said, adding the group also completed "an AI for executives course" and a design session at Stanford University.

That trip left her candid about the gap between markets. "We're way behind in Australia compared to what's happening in Silicon Valley," Doyle said, pointing partly to a cultural reluctance to share knowledge.

"Here we keep secrets in Australia," she said, contrasting it with what she saw at Google: "if you don't share your code and you're not being open, then you're not being a Google citizen."

For HR leaders without an executive travel budget, though, Doyle argued the barrier to entry is lower than it looks. "I would say you don't need a big budget, but you do need a level of curiosity and time," she said, pointing to free vendor workshops, professional-body sessions and simple hands-on experimentation as accessible starting points.

Both leaders were careful to frame AI as augmenting judgment rather than replacing it. "It is just an assistant, it's a coworker," Golani said. "But the human is the ultimate owner, and also the critical thinker – the person who actually has the judgment to understand what is good work from not."

That tension – between AI as collaborator and the anxiety it can generate on the shop floor – echoes recent reporting on staff resistance to poorly explained AI rollouts, and feeds into the wider debate over whether AI could eventually replace entire job categories, rather than simply reshape them.

For both Doyle and Golani, the answer to that anxiety isn't a better tool – it's a better-prepared workforce.