Instrument

AI Adoption Scorecard

A license is not adoption, and usage is not value. Rate your adoption signals across six dimensions and see where you sit on the adoption ladder, which dimensions are holding you back, and the plays to run first. This is the measurement layer of the adoption playbook, the published methodology. Nothing you enter leaves your browser.

1. Reach

Not scored

Of the people who should be using it, how many actively are.

R1The intended-user population is defined and active usage against it is tracked.How to read: You can name who should use it and show who does.
R2Active usage is broad, not concentrated in one team or a few enthusiasts.How to read: Usage is distributed across the roles in scope.
R3Activation is trending up, not flat or decaying after launch.How to read: A usage trend line, not a launch-week spike.
R4Lapsed users are identified and re-engaged.How to read: A defined path for people who stopped using it.

2. Depth

Not scored

How often, and across how many real use cases.

D1Regular users return on a recurring cadence, not one and done.How to read: Repeat-use metrics, not first-use counts.
D2Users apply it across several distinct use cases, not a single task.How to read: A spread of use cases per active user.
D3A maintained library of validated use cases exists for each role.How to read: Persona-based use-case content people can reach.
D4Usage depth is measured, not assumed.How to read: Instrumentation for frequency and breadth of use.

3. Proficiency

Not scored

Whether users are skilled, past basic use.

P1Enablement is role- or persona-based, not one generic training.How to read: Training mapped to what each role does.
P2Users use the tool well, not just at the surface.How to read: Evidence of skilled use, not only logins.
P3Champions or power users support each team.How to read: Named champions with coverage across teams.
P4Capability is assessed, not assumed from attendance.How to read: A skills check, not a sign-in sheet.

4. Workflow Integration

Not scored

Whether it has changed how work gets done, or just been tried.

W1The tool is embedded in standard workflows or procedures, not a side experiment.How to read: It appears in the documented way work is done.
W2It has displaced a slower prior method, not just added a step.How to read: An old step retired, not a new one bolted on.
W3Integration points are owned by the business, not only by IT.How to read: A business owner for each workflow it touches.
W4Friction that blocks daily use is actively removed.How to read: A running list of friction points being cleared.

5. Value Realization

Not scored

Whether measurable outcomes are showing up.

V1Each priority use case has a defined value metric (time, quality, cycle, or cost).How to read: A metric per use case, set before scaling.
V2Outcomes are measured against a baseline, not estimated.How to read: A before number to compare against.
V3The value curve is tracked over time, not claimed once.How to read: The adoption value curve, measured across periods.
V4Realized value is reported to leadership on a cadence.How to read: A recurring value report, not a one-time deck.

6. Governed Adoption

Not scored

Whether usage is happening inside the guardrails.

G1Usage runs on sanctioned tools, and shadow AI is detected and reduced.How to read: A sanctioned-vs-shadow usage picture that is improving.
G2Use is policy-aware: the acceptable-use standard is understood and applied.How to read: Evidence people know and follow the rules of use.
G3Adoption data and governance data are connected.How to read: You can see governed usage, not just usage.
G4Oversight and audit coverage scale with usage.How to read: Controls keep pace as adoption grows.

v1 draft: dimensions, signals, and thresholds are a working model validated per engagement, not a published standard. This playbook is an independent synthesis. It references and maps to recognized frameworks and does not reproduce their content. ADKAR and Prosci are trademarks of Prosci, Inc. PMBOK is a trademark of the Project Management Institute, Inc. The Kirkpatrick Model is a trademark of Kirkpatrick Partners, LLC. The AI Adoption Maturity Model is a work of the Carnegie Mellon University Software Engineering Institute and Accenture. Microsoft, Microsoft 365 Copilot, and Adoption Score are trademarks of the Microsoft group of companies. This work is not affiliated with, endorsed by, or sponsored by any of them.