Calibrate autonomy
Match AI freedom to consequence. Low-consequence drafting may run quickly; high-consequence decisions need explicit human approval.
Pathway 4 · Lesson 0475 min
Coaching People Through AI Change
ADKAR, Kotter, and psychological safety in AI transformation
Learning Objective
By the end of this lesson, you will be able to: Apply the ADKAR model to an AI adoption challenge, design a psychological safety intervention, and create a change management plan for an AI initiative.
75%
of employees are concerned AI will make their jobs obsolete
Engage
"The technology works. The strategy is sound. But people aren't adopting. This is the most common AI transformation failure — and it's entirely predictable. How do you lead people through it?"
Explore
How to apply proven change management frameworks to AI adoption — the human side of AI transformation.
Explain
ADKAR: The Individual Change Model
ADKAR describes the five building blocks every individual needs to successfully change:
A — Awareness: Does the person understand WHY the change is needed? What happens if the organisation doesn't adopt AI? What's the opportunity cost of not changing?
D — Desire: Does the person WANT to change? What's in it for them personally? Fear of job loss is a desire blocker — it must be addressed directly, not avoided.
K — Knowledge: Does the person know HOW to change? Training, frameworks, and coaching provide knowledge. Most organisations jump here first — skipping Awareness and Desire — and wonder why training doesn't stick.
A — Ability: Can the person actually DO it? Practice time, safe environments to experiment, and coaching build ability. Knowledge without ability is just information.
R — Reinforcement: What keeps the change in place? Recognition, celebration of wins, peer accountability, and measuring progress prevent regression.
⚠️
The Most Common ADKAR Mistake
Most AI training programmes jump straight to Knowledge (K) — deploying training without first building Awareness (why this matters) and Desire (what's in it for me). The result: people attend training but don't change behaviour. ADKAR is sequential. Gaps at any stage block all subsequent stages. Diagnose where people are stuck before designing your intervention.
Explain
Psychological Safety: The Foundation of AI Adoption
Amy Edmondson's research on psychological safety is directly applicable to AI adoption. People won't experiment with AI if they fear:
Five psychological safety signals leaders must send:
1. Safe to fail: 'Experiment with AI. If it doesn't work, that's learning, not failure.'
2. Safe to ask: 'There are no stupid questions about AI.'
3. Safe to challenge: 'Push back on AI adoption decisions you disagree with.'
4. Safe to not know: 'I don't understand everything about AI either.'
5. Safe to experiment: 'You have dedicated time to explore AI tools.'
Elaborate · Worked Example
Case Study: Microsoft Copilot — People-First Change
Context
Microsoft's internal rollout of Copilot to 200,000+ employees.
The Situation
The challenge: Rolling out AI to a workforce with widely varying AI experience, significant anxiety about job displacement, and a culture that values technical expertise. The approach: - Awareness: CEO and leadership communicated the 'why' — competitive necessity and personal opportunity — before any tools were deployed - Desire: Framed as 'AI makes you better at your job' not 'AI replaces your job' - Knowledge: 5,000 early adopters trained as peer coaches before company-wide launch - Ability: Dedicated 'AI experimentation time' — 2 hours per week for 6 weeks - Reinforcement: Weekly wins shared in team meetings, manager-led adoption modelling
The Outcome
90% active weekly usage. 3.5× higher adoption in manager-led teams. 6 weeks to habitual use vs. 16-week industry average. The technology was identical to competitors — the change management was the differentiator.
Try It · Reflect & Apply
ADKAR diagnosis: Choose an AI initiative where adoption is lower than expected. For each ADKAR stage, rate the current state (1-5) and identify one specific intervention. Where is the biggest gap?
Your notes are saved in your browser for this session.
Evaluate · Knowledge Check
1 / 3Why do most AI training programmes fail to change behaviour?
External video references
Keep Learning on YouTube
Open in a new tab
Architect Leadership Lab
Change Leadership Lab
Leadership decision
Identify the ADKAR stage where your most important stakeholder group is currently stuck and choose one intervention.
Evidence to create
A stakeholder heat map naming the audience, perceived loss, desired gain, current ADKAR stage, sponsor, and next conversation.
Meeting retrieval prompt
Ask: ‘What are people afraid they will lose if this AI initiative succeeds—and have we spoken about it plainly?’
Spaced application: Return to this decision after 7 days, then after 30 days. Record what evidence changed your initial assumption and what you will do next.
Curated external references
Executive Resource Library
Select one this week
Leading AI change managementExternal practitioner sessionA practical overview of leading AI adoption through communication, support, and implementation discipline.Watch · 30 min WorkLab: The PodcastMicrosoft WorkLabA curated series on how AI is reshaping roles, careers, operating models, and workforce choices.Watch · Choose 1 episode ADKAR overviewProsciOptional reinforcement on the individual change model used to diagnose adoption barriers.Read · 10 min
Human-centred work design
Human Oversight Check
Use this original practice tool to test whether an AI-enabled workflow supports real human judgment, rather than merely placing a person at the end of an automated process.
Self-check
0/5
Start by making the work and decision rights explicit
Keep context visible
Make the objective, inputs, constraints, source material, and current system state easy to inspect before anyone approves an output.
Design active review
Review should require informed judgment, a plausible override, and the authority to act—not a symbolic sign-off after the fact.
Route exceptions clearly
Specify what uncertainty, conflict, missing evidence, or risk looks like—and who receives the escalation with enough context to decide.
Maintain human capability
Protect deliberate practice in the work that people may one day need to verify, challenge, or perform without AI assistance.
Apply it to one workflow
Ready to move on?
Mark this lesson complete and continue to "Leading as the Orchestrator".