Most organisations approaching AI do something predictable: they identify the technology, assess the features, build a roadmap, and allocate a budget. The strategy looks complete. The logic is sound. And then, six months later, very little has actually changed.
The problem isn't the strategy. It's that the strategy was built on a simplified picture of where the organisation is — and the simplification cut away exactly the things that determine whether any of this works in practice.
When an organisation decides to adopt AI, the conversation typically starts at the level of tools — which platform, which vendor, which use cases. What doesn't get asked is what the organisation actually looks like at the level of the work being done.
Not the org chart. Not the process documentation. The actual work — what people do, how they do it, where the friction is, which processes have been running on institutional knowledge for years, and which roles are already halfway to something different without anyone having named it yet.
That inventory doesn't exist in most organisations. And without it, an AI strategy is essentially a plan built on assumptions about a system nobody has actually looked at closely.
There is a specific kind of distance that appears in every organisation navigating this shift — the distance between what leadership sees and what is actually happening at the level of delivery.
A team lead knows things about their team's work that no executive dashboard captures. They know which processes could be automated today and which ones look simple but aren't. They know which people are already using AI informally and which ones are quietly worried about what it means for them.
The organisations that navigate AI adoption well start by closing that distance — not by cascading a strategy downward, but by building the picture from where the work actually happens.
The most useful frame for thinking about AI in delivery isn't a maturity model or a capability assessment. It's a simple question applied to each process, each step, each handoff: is this assisted, automated, or zero-touch? And what would it take to move it one step forward?
That question, asked seriously and answered honestly, produces something a strategy document never does: a real picture of where the organisation is, what's possible in the short term, and what needs to change before the longer term becomes accessible.
It also produces something else — the conversation about what happens to the roles in between. Which is the conversation most organisations are most reluctant to have, and the one that determines whether any of this lands well or lands badly.
Every organisation is asking which AI tools to use. Almost none are asking what their organisation looks like in three years if they get this right — and building toward that picture deliberately, rather than arriving at it by accident.
That's not a technology question. It's the question that has to come first — before the tools, before the roadmap, before anything gets simplified into a strategy that looks right on paper.