The organisations winning with AI right now have one thing in common, and it has nothing to do with their tech stack.

There's a moment happening in enterprise AI integration that most leadership teams are going to miss.

Not because the technology isn't ready. It is.

Because they're asking the wrong question.

The question I keep hearing from boards and executive teams is: "What AI tools should we be using?"

The question that actually matters is: "Do our people know where the bottlenecks are, and do they trust us enough to tell us?"

I've spent twenty years helping organisations answer that second question. The technology changes. The pattern doesn't.

The shift

Microsoft just announced a capability inside Copilot called "Co-Work", an agentic AI layer that runs background workflows, automates repeatable tasks through natural language, and centralises outputs in a single command view.

It's not a new product. It's an evolution of a platform millions of organisations are already paying for. Anthropic's Claude is operating in the same space. So is Google's Gemini. The tooling is converging fast.

What this means is simple: the technology barrier to AI adoption, the one most organisations have been pointing to, is about to disappear. Within 12 months, the average enterprise employee will have agentic AI capability inside the tools/tasks they already use every day.

And that's where the real challenge begins.

Technology was never the hard part

I've been working with organisations on transformation for over two decades, originally through culture diagnostics, leadership development, and what we now call psychosocial risk. In the last twelve months, I've been building AI-integrated operational systems: safety command centres, automated risk dashboards, fatigue intelligence models, scenario planning tools.

Different tools. Same lesson.

The technology works. The models are capable. The integrations are solvable.

What makes or breaks adoption is culture. Every time.

It comes down to three things.

Whether people feel safe enough to name the real problems.

The most valuable input for any AI integration isn't a process map drawn in a boardroom.

It's the honest answer from a supervisor who says: "I spend four hours every week manually copying data between systems because no one ever connected them."

Or the project manager who admits: "We've been running this report by hand for two years because the automated version broke and no one had the authority to fix it."

These are the bottlenecks AI can eliminate, overnight. But they only surface when people trust that honesty leads to improvement, not blame.

Whether leaders are willing to let the problem definition come from the ground up.

Most organisational strategies are designed top-down, Ai integration is no exception. A consulting firm/ executive teams map processes, identify "quick wins," and present a “transformation roadmap” for on the ground implementation.

The problem is that the people who actually do the work were never in the room. The quick wins don't match the real friction points. The roadmap looks impressive in a board pack and dies in execution.

The organisations I've seen get genuine traction flip this entirely. They run a simple exercise: What are the most time-consuming tasks in your day? Where do decisions get stuck? What would "automated in the background" look like for you?

Then they map those answers into 30–90 day priorities. The AI strategy becomes something people built together, not something done to them.

Whether the goal is efficiency alone, or efficiency plus engagement.

This is the distinction most organisations miss.

AI can deliver operational cost efficiencies, that's well established. But the window right now is to deliver those efficiencies at the same time as a meaningful shift in how people experience their work.

When someone sees that the repetitive, low-value task they've been doing for years is suddenly automated, and that their input was the reason it happened, something changes. They stop seeing AI as a threat and start seeing it as something that makes their role better.

That's not a technology outcome. That's a culture change. And in twenty years of doing this work, it's the single most reliable predictor of whether a transformation sticks.

The window

Right now, AI is moving from something early adopters experiment with to something embedded in the platforms everyone already uses.

The organisations that have done the groundwork, mapped their bottlenecks, engaged their teams, built short-cycle rollout plans, will plug into these capabilities and see value in weeks. Soon, that will become days.

The organisations that haven't will do what they've always done: form a committee, hire a consultant to tell them what they already know, and watch twelve months pass before anything changes.

The difference isn't budget or technical capability. It's whether someone had the foresight to start the conversation with their people before the platform forces the decision.

What this looks like in practice

I recently worked with an organisation in heavy industry, crane operations, safety compliance, roster management. Not exactly the first sector you'd think of for AI transformation.

Within 3 days, we had a working risk command centre: automated gap analysis across eight compliance domains, scenario modelling with real cost exposure data, draft communications generated for three levels of the organisation, and a fatigue intelligence dashboard connected to live roster patterns.

None of that required cutting-edge research. It required someone sitting down with the safety manager and asking three questions: Where does information get stuck? What takes you the longest? What would you want on a dashboard if you could have anything? What are the biggest challenges in your role, how do they impact your teams/ organisation?

The AI did the building. The human did the thinking. That's the model that works, and it's the same model that's worked for me across two decades of transformation work, long before AI entered the picture.

The question worth asking

If you're in a leadership role thinking about AI strategy, here's where I'd start, and it has nothing to do with technology:

If you asked every team in your business to name the three things that slow them down the most, would they feel safe enough to give you an honest answer?

If yes, you're ready. The tools will meet you where you are. Get them in a room and have those conversations.

If not, that's the work that matters most. And it's work no AI platform can do for you.

Kim Yabsley is the founder of Archispeaks, a consulting practice that turns culture intelligence into operational performance. For twenty years she has helped organisations identify what's working, do more of it, and build the conditions where collaboration, innovation, and efficiency happen together.