Learning objective
Distinguish Awareness, Desire, Knowledge, Ability and Reinforcement, then choose support for the earliest evidence-backed adoption barrier in a specific group.
Why it matters
Change activity is not change adoption. A launch can achieve 95% training attendance while people still cannot explain the purpose, reject an unfair trade-off, fail realistic exceptions or return to the old workflow because managers continue to reward it.
Core concept: the earliest barrier
ADKAR treats individual adoption as five outcomes:
The sequence helps prioritise diagnosis. It is not a one-time waterfall: later outcomes can weaken, and a change in workflow or incentives can create a new earlier barrier.
Scenario
A hospital introduces a new discharge workflow. Nurses can explain the patient-safety reason and support the goal. They pass the e-learning assessment, but in a busy shift they cannot complete the medication-reconciliation step because the required pharmacy data arrive late.
Worked decision
The earliest supported diagnosis is Ability, not Knowledge. People know the procedure and are trying to use it, but the real environment blocks performance. The response is to repair the data hand-off, rehearse realistic cases and provide escalation—not repeat the course.
Decision rule
Choose the first outcome for which evidence is insufficient. Treat it as a revisable hypothesis about a segment and context, never as a personal performance label.
Ethical boundary
Low Desire may be a rational response to workload, harm, status loss or lack of voice. Diagnose the design and decision rights before trying to “overcome resistance.” Keep personal comments confidential and report only useful aggregation.
AI lens
AI may group authorised questions or simulate routine practice. It must not infer motivation, loyalty or emotion, assign employee Desire scores, or suppress dissent.
Quick check
A user passes training but cannot complete the task with the available permissions. Which outcome is the likely barrier?
A. Awareness.
B. Desire.
C. Knowledge.
D. Ability.
Answer: D. The user knows what to do but lacks the real conditions to perform it.
Next step
Use the ADKAR workspace to diagnose the field-service AI adoption case, then compare the model with the distributed-change analysis in Fiction Lab.