Most organisations I speak with know something is wrong before they call me.
They know they have a problem. They can see the symptoms. Decisions that should take days are taking weeks. Teams that should work together have built workarounds and shortcuts to avoid each other. Initiatives keep launching but improvement keeps failing to materialise. Somewhere along the way, a collective resignation has set in. There is the assumption that this is just how large, complex organisations work.
It often is, but it doesn’t have to be. The path out is not what most people expect.
The instinct, almost universally, is to reach for a solution. Organisations will try a restructure, a new tool or technology, a revised process, or a full transformation programme. And it makes sense. Implementing solutions feels like progress. Diagnosis feels like delay.
That framing is exactly backwards.
Why the diagnostic has to come first
An operating model problem is rarely what it looks like on the surface. When I see a team that cannot execute, I don’t assume the team is the problem. When I see a process that keeps breaking down, I don’t assume the process needs fixing. I start from the assumption that operating models are systems, which means that symptoms and causes are usually in different places.
A team struggling to deliver may be working with a mandate that conflicts with their actual authority. A process that keeps failing may be the result of conflicting incentives across teams. A leadership team that cannot agree on priorities may be operating without a clear enough definition of what the business is actually trying to do.
None of those things show up in an activity log or a project post-mortem. They show up when you know what to look for, and when you’re willing to look in the right places.
What the methodology actually looks like
A diagnostic is not a survey. It is not a series of stakeholder interviews that get summarised into a slide deck. Done properly, it is a structured process of evidence gathering, pattern recognition, and system mapping.
In practice, it tends to move through three phases.
The first is orientation: understanding the operating context well enough to know where the real diagnostic work needs to happen. This means getting clear on strategy, on what the organisation is trying to achieve, and on where the significant points of friction are. It is deliberately broad, because you are looking for signals before you start narrowing.
The second phase is where most of the substantive work happens. By talking to the people who actually do the work, in addition to the leaders who are experiencing the impact, I’m examining how the organisation actually functions. I review how decisions get made, where accountability is clear and where it dissolves, how resources flow relative to stated priorities, how different parts of the organisation interact in practice versus in theory. I’m looking at the seven signs I’ve written about elsewhere (and will write more about next week), but more importantly I’m looking at the relationships between these signs. Which problems are reinforcing each other? Where is the root cause?
The third phase is synthesis. This is where pattern recognition is critical for accurate diagnosis. I produce not just a list of findings, but a coherent account of what is actually going on in the system, why it is producing the outcomes it is producing, and what that means for what needs to change.
The output of a diagnostic is not a set of recommendations. The value of a diagnostic is clarity. A diagnostic provides a shared understanding of what the real problem is, specific enough to make good decisions about what to do next.
The AI dimension
I want to say something about this directly, because it is increasingly the context in which these conversations are happening.
A lot of organisations are investing heavily in AI right now. Agentic tools, automation, intelligent workflows. The technology is genuinely powerful and the opportunity is real. But I keep seeing the same pattern: organisations deploying AI onto operating models that are already broken, and then being surprised when the results are underwhelming.
AI does not fix structural misalignment. It does not resolve unclear accountability. It does not compensate for a strategy that has never been properly translated into operating terms. What it does is amplify what is already there. Which means that if what is already there is dysfunctional, AI makes it dysfunctional faster.
This is not an argument against AI investment. It is an argument for knowing what you are working with before you build on top of it. The diagnostic is not a detour from the AI agenda. For most organisations, it is the prerequisite to getting a meaningful return from it.
The uncomfortable truth
Most organisations resist the diagnostic because it takes time they feel they don’t have. The irony is that every initiative they launch without it takes longer, costs more, and delivers less than it should. The diagnosis feels like a delay. In practice, it is what makes everything else faster.
If you’re sitting with a nagging sense that your organisation is working harder than the results justify, that’s worth paying attention to. It usually means the operating model has something to tell you.
Next, I’ll go deeper on the seven signs including the one that organisations are most likely to misread.