You've Got the Licences. So Why Isn't Anyone Using AI?

A pattern is emerging across Australian organisations that have made the leap to AI tools, particularly Microsoft Copilot. Leadership is on board. IT has rolled out licences. There may even have been a training session. And yet, six months later, most staff are still writing emails and searching Google the same way they always have.

This is not an AI problem. It is an adoption problem. And it is almost universal.

The licence trap

In conversations with organisations across the property, construction, marketing, and professional services sectors, we hear the same thing repeatedly. A company invests in an enterprise AI platform, delivers a general overview session for staff, and then waits for productivity to improve.

It rarely does.

The assumption baked into this approach is that AI adoption works like software adoption: give people access and a walkthrough, and they will figure it out. But AI does not behave like software. It does not have a set of buttons to learn or a workflow that stays the same. It requires something much harder to change: behaviour.

People do not struggle with AI because they cannot find the right menu. They struggle because they do not know how to think with it, how to prompt it effectively, or how to trust what it produces. That is a different problem entirely, and a one-hour webinar will not solve it.

What "good prompting" actually means

A useful illustration: when we recently spoke with an HR leader at a major WA government development body, she shared something that stuck. She and a colleague had developed a real fluency with Copilot. Their prompts were detailed, contextual, and specific. Meanwhile, staff around them were typing things like "write me my executive paper on this" and wondering why the output was disappointing.

The gap was not access. Everyone had the same licence, the same tool, the same potential. The gap was understanding what AI actually needs from you to do useful work.

Effective prompting is not a technical skill. It is a communication skill. It involves providing context, specifying the audience, defining the format, and being clear about what you want. Most people have never been taught this. They use AI the way they use a search engine and then conclude that it does not work.

This is where structured, role-specific training changes everything. Once someone understands how to work with AI in the context of their actual job, they tend to get it quickly. The "aha moment," as we have come to call it, happens at the role level, not the platform level.

Role-based training is not a nice-to-have

Generic training has its place. A broad awareness session can help an organisation establish a shared baseline and reduce fear. But it is not sufficient on its own, and too many organisations treat it as if it is.

The reason role-specific training works is straightforward. A project manager and a marketing coordinator and an executive assistant all use AI very differently. They have different workflows, different outputs, different risks, and different opportunities. When training is tailored to their actual day-to-day, they can immediately see how to apply what they have just learned. When it is not, they leave the session with general knowledge that never quite connects to the work in front of them.

Across every engagement we have run in the past year, whether with property developers, creative agencies, or commercial construction businesses, this has held true. Organisations that run role-specific sessions report higher confidence, faster adoption, and more consistent use of AI tools beyond the training period. Those that rely on a one-size-fits-all approach rarely sustain the momentum.

The workforce confidence problem is real, and it is manageable

There is a particular dynamic in organisations with a wide age range in their staff. Younger employees often arrive with some existing fluency. They may have used AI personally, have fewer inhibitions about experimenting, and are quicker to integrate it into their work. Longer-tenured staff, particularly those 50 and above, often bring a different set of concerns.

Some of those concerns are practical: they have not been trained, they are uncertain what is safe to put into the tool, and they worry about producing poor-quality work. Others are more existential: a fear that demonstrating how efficient AI makes them could work against their job security.

Both are understandable. Neither is insurmountable.

What resolves them is not reassurance. It is capability. Once someone has had the experience of using AI effectively, the anxiety typically drops. Seeing concrete examples from colleagues doing similar work is more convincing than any messaging about AI strategy from the top of the organisation. Peer demonstration, structured experimentation time, and a training environment where it is acceptable to fail and learn are more effective than most organisations give credit for.

Governance is not bureaucracy. It is a confidence enabler.

One of the most consistent barriers to adoption is not resistance to AI itself. It is uncertainty about what is permitted. When people do not know whether they can put client information into an AI tool, whether their outputs will be auditable, or whether their organisation's data might end up somewhere unintended, they default to doing nothing.

A clear, practical AI usage policy — one that specifies what platforms are approved, what categories of information are safe to use, and what the boundaries are — removes that ambiguity. It does not make AI safer. It makes people confident that it is safe, which is what actually drives use.

Governance documents that simply say "you can use Copilot, but nothing confidential" without defining what confidential means, or what the approved use cases are, leave too much open to interpretation. People will interpret ambiguity conservatively. That means they will not use the tools.

Getting the governance layer right is foundational. It should come before the training, and it should be written in plain language that operations and frontline staff can actually act on.

What structured adoption actually looks like

Based on our work across the past two years, organisations that build genuine AI capability tend to move through three stages, even if the timelines vary.

The first is shared foundations. Everyone in the organisation develops a working understanding of what AI can and cannot do, how to prompt effectively, and what the governance guardrails are. This does not need to be deep. It needs to be consistent.

The second is role-based capability. Teams learn to apply AI to their specific workflows. This is where the practical value becomes visible and where adoption starts to become self-sustaining. People start sharing what is working. They bring AI into their daily routines because it is genuinely saving them time.

The third is ongoing integration. AI capabilities evolve quickly, and so do the tools. What was true of Copilot six months ago is not entirely true today. Organisations that build a culture of continuous learning — through peer sharing, periodic refreshers, and internal champions — are able to keep pace with change rather than falling behind every time a new feature drops.

The organisations that struggle are those that treat AI adoption as a project with a beginning and an end. It is not. It is a capability that requires ongoing investment, like any other.

The question worth asking

If you have AI licences deployed across your organisation and are not seeing meaningful adoption, the problem is almost certainly not the technology. It is the absence of a structured, behaviour-change-focused approach to helping people integrate AI into how they actually work.

The organisations making real progress are not the ones with the most sophisticated tools. They are the ones that have taken the training seriously, been deliberate about governance, and invested in making AI feel accessible and relevant to the people doing the work.

That is where the competitive advantage will be won. Not in the technology itself, but in how well your people learn to use it.

If your organisation has AI tools deployed but adoption is lagging, we would be glad to help you understand why and what to do about it.