Here’s a take that might ruffle some feathers: the biggest barrier to AI adoption in Australian organisations isn’t the C-suite and it isn’t frontline workers. It’s the people in between — middle managers — and almost nobody is giving them the support they actually need.

I’ve spent the better part of 15 years working in learning and development, and the pattern is unmistakable. Executive leadership buys the vision. Frontline staff are often curious and willing. But middle managers? They’re caught in a vice, squeezed between top-down mandates to “be more innovative” and bottom-up anxiety about what AI means for jobs. And instead of helping them navigate that, most organisations just pile on more pressure.

The Squeeze Is Real

Think about what we’re asking middle managers to do right now. Implement new AI tools they may not fully understand. Reassure their teams that nobody’s getting replaced. Hit the same KPIs they’ve always had. And do it all while the ground shifts underneath them.

A 2025 workforce survey from AHRI found that while 68% of Australian senior executives said they were confident about AI integration timelines, only 31% of middle managers felt the same way. That gap is enormous, and it tells you everything about where the real friction lives.

Middle managers aren’t resistant because they’re stubborn. They’re resistant because they’re rational. They can see the implementation challenges that the boardroom can’t. They know which team members will struggle, which workflows will break, and which “quick wins” are anything but. And most of the time, nobody’s asking them about any of it.

Why Organisations Keep Getting This Wrong

There’s a standard playbook for AI adoption that goes something like this: CEO announces AI strategy, IT department selects vendors, training team runs a couple of half-day workshops, and then everyone wonders why adoption stalls at 20%.

The workshops are almost always targeted at either executives (“here’s the strategic vision”) or end users (“here’s which buttons to click”). Middle managers get a bit of both and a lot of neither. They’re expected to be change agents without being given the tools, time, or authority to actually manage change.

I worked with a government department in Canberra last year where this played out almost comically. They rolled out an AI-powered document review system. The executive team loved it. The admin staff were trained on it. But the team leaders — the people who actually had to reorganise workflows, manage the transition period, and deal with the inevitable teething problems — got a 90-minute briefing and a FAQ document.

Three months later, half the teams had quietly gone back to the old system. Not because the technology was bad, but because nobody had supported the people responsible for making it stick.

What Middle Managers Actually Need

Based on what I’ve seen work — and what I’ve seen fail — here’s what organisations should be doing differently.

Give them a genuine voice in the planning process. Middle managers know things the C-suite doesn’t. They know which processes are actually ripe for automation and which ones look simple but are riddled with exceptions and edge cases. Bring them in during the scoping phase, not after decisions have been made.

Invest in their AI literacy — properly. Not a two-hour overview. Not a webinar. I’m talking about sustained, practical learning that helps them understand what AI can and can’t do in their specific context. They don’t need to become data scientists, but they do need enough knowledge to ask good questions and make informed decisions about where AI fits in their team’s work.

Reduce the competing demands during transitions. This is the one that never happens but always should. If you’re asking a manager to lead an AI rollout on top of their normal workload, something has to give. Either adjust their targets, bring in temporary support, or extend the timeline. Pretending it can all be done with the same resources is a recipe for burnout and failure.

Create peer networks. Some of the most effective AI adoption I’ve seen happened when middle managers could learn from each other. Not in a formal training setting, but in small groups where they could share what was working, what wasn’t, and how they’d solved problems. It sounds simple because it is. But most organisations don’t create the space for it.

The Bigger Picture

Australian businesses are spending billions on AI. The Australian Computer Society estimates that AI-related investment will exceed $5 billion annually by 2027. But spending money on technology without investing in the people who have to implement it is like buying a Formula 1 car and handing the keys to someone who’s only driven automatics.

Middle managers are the connective tissue of any organisation. They translate strategy into action, manage competing priorities, and hold teams together through change. If we’re serious about AI adoption — and the economic stakes suggest we should be — then we need to stop treating them as an afterthought.

The organisations that get this right will be the ones that move fastest. Not because they bought the best technology, but because they invested in the people who make technology work in practice.

It’s not a glamorous insight. There’s no shiny product to sell, no transformation framework to licence. Just the hard, unglamorous work of supporting the people in the middle. But from where I sit, it’s the single highest-impact thing most Australian organisations could do right now.