I have run enough change management programmes to know what most of them look like: a communications plan, a stakeholder matrix, some training sessions, and a set of adoption metrics. They work, technically. People comply. But compliance is not the same as capability, and when the change is AI adoption, the gap between compliance and capability is where the value dies. Josiah has given me a different perspective: what if the change management strategy was a learning design?
What You Need to Know
- Standard change management produces compliance-level AI adoption. Learning design produces capability-level adoption that compounds value over time.
- The key difference: change management provides information. Learning design provides scaffolded practice with feedback.
- Replace the communications plan with a learning journey: foundational understanding, guided practice, independent application, peer sharing, and extended practice
- The success metric should be "percentage of people who can independently apply AI to their work," not "percentage using the tool"
The Compliance Problem
Standard change management aims for adoption. People use the new system. They follow the new process. They attend the training. The metrics look good.
But when the change is AI, compliance-level adoption produces compliance-level outcomes. People use the AI tool the way they were told to, in the specific scenarios covered by training, with the exact prompts they were taught. They don't experiment. They don't innovate. They don't discover new applications. They comply.
This matters because AI's value scales with creative application. An AI tool used in exactly the way training demonstrated is delivering a fraction of its potential. An AI tool used by people who understand it well enough to adapt it to new situations delivers compounding value.
In AI adoption, surface learners use the tool as trained. Deep learners use the tool as needed, including in ways nobody anticipated. Both adopted the tool. Only one group is building capability.
Dr Josiah Koh
Education & AI Innovation
What Learning Design Brings
Scaffolded Capability Building
Change management typically introduces the new system all at once: here's the tool, here's how it works, here's the training. Learning design scaffolds the introduction: here's one capability, practise it until comfortable, then here's the next one.
Scaffolding respects the learning curve. People master the basics before encountering complexity. Confidence builds alongside capability. The result is not just adoption but genuine proficiency.
Practice With Feedback
Change management provides information. Learning design provides practice with feedback. The difference is significant: practising a skill with feedback builds capability in ways that watching a demonstration or reading a guide cannot.
For AI adoption, this means hands-on sessions where people use the tool on real tasks and receive feedback on their approach. Not "did you get the right answer?" but "was your approach effective? What could you try differently?"
Assessment That Drives Improvement
Change management measures adoption rates. Learning design measures comprehension, application, and adaptation. These measurements don't just report progress. They identify where people are struggling and what support they need.
When assessment reveals that 70% of a team understands the AI tool conceptually but only 30% can apply it to their workflow, the response is specific: more hands-on practice with workflow-specific scenarios. Change management's response to the same data would be: more communication.
Making It Work
Replace the Communications Plan With a Learning Journey
Instead of a 12-week communications schedule (week 1: announcement, week 3: demo, week 6: training, week 8: go-live), design a learning journey:
- Weeks 1-2: Foundational understanding (what AI is, what it can do, what it means for you)
- Weeks 3-4: Guided practice on simple tasks (with feedback and support)
- Weeks 5-6: Independent practice on real work tasks (with coaching available)
- Weeks 7-8: Peer sharing and collaborative problem-solving
- Weeks 9-12: Extended practice with increasing complexity and decreasing support
Use Champions as Coaches, Not Advocates
Change management champions advocate for the change. Learning design champions coach their peers through the learning process. The distinction matters: advocacy persuades, coaching enables. People who are enabled to succeed don't need to be persuaded.
Measure Capability, Not Adoption
The success metric is not "percentage of people using the AI tool." It is "percentage of people who can independently apply AI to their work context." The first metric rewards compliance. The second rewards learning.
Change management and learning design are not competing approaches. They're different lenses on the same challenge. But when the change is AI, the learning lens produces deeper, more durable, and more valuable outcomes than the change management lens alone. The question is not "how do we get people to use AI?" It is "how do we build people's capability to work with AI?" The answer to the second question solves the first.