measurement
The Institutional AI Readiness Model
Organizations are never at one stage. The gap between their highest and lowest dimension predicts failure better than their average does.
Structure
- Experimenting
- Activity exists, uncoordinated, individually driven.
- Building
- Deliberate capability is being assembled; little of it is in production yet.
- Scaling
- Multiple capabilities are in production and the operating model is straining.
- Transforming
- AI is a normal part of how work is designed and governed.
No organization sits at a single stage. The ten dimensions, strategy, leadership and ownership, data, architecture, governance, security and identity, workflow readiness, workforce and adoption, delivery capability, and measurement, each move at their own pace, and each can be scored independently across four stages: experimenting, building, scaling, and transforming.
The average score is the least useful number this model produces. An organization scaling on strategy while still experimenting on governance is not "building" on average; it is carrying a governance gap into production at scale, which tends to surface as an incident rather than a gradual correction. The gap between an organization's highest and lowest scoring dimension predicts where the next failure comes from far better than the average across all ten does.
Reading the model honestly means naming the lowest dimension out loud, and treating it as the actual constraint on how fast the rest of the organization is allowed to move.