Why this matters
Across the organizations we meet, the AI conversation usually begins with a tool question: which platform, which model, which vendor. It is rarely the right first question. The organizations that get value from AI are not the ones that moved fastest on procurement — they are the ones that understood their own work clearly enough to know where help was needed.
Readiness is commonly framed as an infrastructure question: do we have the data, the systems, the connectivity. Those matter. But in practice, the first constraints are human: nobody has agreed what problem to solve, the workflow as documented does not match the workflow as lived, and the people closest to the work have not been asked what actually slows them down.
What we are seeing
In structured readiness assessments, the same pattern repeats. Capability scores are middling but fixable. Data quality is uneven but rarely impossible. The real spread appears in two places: governance — who is allowed to decide, and what must stay human — and shared understanding, where the leader's picture of the workflow differs sharply from the operator's.
That gap is not a problem to hide. It is the most useful diagnostic signal available, because disagreement between roles almost always marks the spot where a workflow is held together by memory and goodwill rather than by design.
Practical implications
A readiness review that earns its cost produces three things: a prioritised list of use cases tied to real work, a gap list that names what must be fixed before scale, and a shared language the whole leadership team can argue with. If an assessment produces only a score, it has told you very little.
- Treat different answers between roles as findings, not failures.
- Score governance and process alongside data and tools.
- End every assessment with two candidate pilots, not forty ideas.
Actions you can take
- Run one honest readiness review with both a leader and an operator in the room.
- Write down your top three candidate AI use cases and the workflow each one touches.
- Name one decision that must always stay with a person, before you automate anything.
Limits — what stays uncertain
- These observations come from our own assessments and community sessions, not a continental survey.
- Readiness frameworks evolve; treat any score as a conversation starter, not a certification.

