ARM Hub Director Programs & Commercial Corrie Germin has spent 20-plus years helping businesses and governments decide how to invest in new technology. He now leads programs at ARM Hub designed to help businesses stand up AI Adoption projects. His advice is simple: one day at a time.

By late 2026, the big question for an AI project isn't whether the technology works. It's whether the organisation can actually handle using it. Germin learned this while working on a digital identity project for a major federal government department. "Technology transformation was the easy bit," Germin said. "Changing culture internally was the biggest challenge."

THE CONCERN: "Our last pilot went nowhere."

THE EVIDENCE SAYS: Pilots stall when the problem is too broad, expectations aren't agreed on, or the team isn't ready to change how it works.

THE TAKE AWAY: Run the project in stages, with go/no-go checkpoints agreed before work starts.

Read more

What makes a project succeed

Manufacturers face supply chain and geopolitical risks that make investment decisions harder, so the process behind a project matters as much as the tech itself.

That's important, and it's why Germin pushes for staged projects with clear "go/no-go checkpoints" at each phase: feasibility, R&D, then construction.

He says the best projects usually share three things in common:

• a clear purpose,

• simple integration,

• a real benefit to people.

Projects that solve problems like workplace health and safety tend to get prioritised first, and can pay back their upfront cost in under three years.

Getting started

Before anything else, Germin says businesses need to ask basic questions: how accurate is your data? Do you trust your dashboards? How many different software systems are you juggling?

Beyond that, it comes down to culture, whether the team is actually willing to change how it works to fit the new tool. Germin points to an agri-tech company that tried replacing manual labour with robots. The tech worked fine, but the company hadn't trained enough staff to run and manage it, so the rollout stalled.

Knowing when to stop

Sometimes the right call is to pull the plug. If a team and client never agreed on what "success" looks like at the start, the project can drift until nobody's on the same page.

"If not agreed, expectations take over and both the delivery team and client are unhappy," Germin said. "This is when you pull the plug."

Germin is also upfront that early-stage AI projects come with real uncertainty. An early prototype phase might only cover 25% of the total cost, it's not until later that a business can be more than 90% confident about the full price tag.

The numbers

Australia's AI adoption numbers vary depending on who's counting. The ABS found 12% of businesses used AI in 2024–25 (35% of large businesses, down to 11% of small ones). A separate National AI Centre survey put SME adoption at 43%; while Deloitte found 28% of Australian businesses had moved at least 40% of their AI pilots into full production, with 61% seeing gains in efficiency. Tellingly, more than half said skills gaps were still a major barrier.

Businesses are experimenting and seeing some wins. Turning that into reliable, everyday operations is the next step. Get in touch to find out more about how to get started.