I've been assessing enterprise AI maturity for two years now. Every organisation I work with has a technology maturity assessment. Almost none have a change maturity assessment. Which is odd, given that the evidence consistently shows that change capability is a better predictor of AI success than technology capability.
What You Need to Know
- AI change maturity describes how well your organisation manages the human side of AI adoption
- Most enterprises are at Level 1 (Ad Hoc) or Level 2 (Reactive), even if their technology maturity is advanced
- Each level has specific characteristics, risks, and actions to progress
- Organisations at Level 3+ adopt AI 2-3x faster and sustain adoption longer
The Five Levels
Level 1: Ad Hoc
Characteristics: No structured approach to AI change management. Individual teams handle adoption however they see fit. Some teams succeed through local initiative. Others struggle with resistance and low usage. No consistent metrics.
Risk: High variation in adoption across the organisation. Successful AI use is dependent on individual champions rather than organisational capability. If a champion leaves, adoption collapses.
How to progress: Appoint a single person responsible for AI adoption across the organisation. Define basic adoption metrics. Start measuring.
Level 2: Reactive
Characteristics: Change management exists but responds to problems after they surface. When a team reports low adoption, someone investigates. Training programmes exist but are generic. Communication happens but isn't targeted.
Risk: Always playing catch-up. By the time resistance is visible, it's entrenched. The organisation spends more time fixing adoption problems than preventing them.
How to progress: Shift from reactive to proactive. Assess readiness before deployment, not after. Build a change management workstream that runs in parallel with technical delivery.
Level 3: Structured
Characteristics: A defined change management methodology is applied to every AI deployment. Readiness assessments happen before build. Communication plans are targeted by stakeholder group. Champions are selected and supported. Adoption metrics are tracked weekly.
Risk: The methodology becomes rigid. Different teams need different approaches, and a one-size-fits-all framework misses the nuance. There's also a risk of change management becoming a bureaucratic overlay rather than an enabler.
How to progress: Build flexibility into the framework. Adapt the approach for different team contexts while maintaining core principles. Develop internal change management capability beyond a single team or person.
Level 3 is where most organisations should aim to be. It's the point where change management stops being a project activity and becomes an organisational capability.
Tim Hatherley-Greene
Chief Operating Officer
Level 4: Integrated
Characteristics: Change management is embedded in how the organisation operates, not a separate function. Every AI initiative includes change management as a default, not an add-on. Leaders at all levels understand their role in supporting adoption. The organisation learns from each deployment and improves.
Risk: Complacency. "We're good at this" can become "we don't need to keep investing." Change capability, like any capability, degrades without maintenance.
How to progress: Institutionalise learning. Build feedback loops from every deployment into the change methodology. Connect adoption outcomes to leadership performance metrics.
Level 5: Adaptive
Characteristics: The organisation's change capability evolves as fast as its technology capability. Change approaches are customised in real time based on team-level data. Resistance is anticipated and pre-empted. Adoption is measured continuously and optimised like a product metric.
Risk: Over-engineering. Not every deployment needs a sophisticated change programme. Maintaining this level requires investment that may not be justified for routine deployments.
This level is rare. It's the aspiration, not the immediate target.
Where Most Organisations Are
Based on the assessments I've conducted, the distribution looks roughly like this:
- Level 1 (Ad Hoc): 40% of enterprises
- Level 2 (Reactive): 35% of enterprises
- Level 3 (Structured): 20% of enterprises
- Level 4-5 (Integrated/Adaptive): 5% of enterprises
The correlation with AI adoption success is clear. Organisations at Level 3+ report significantly higher sustained usage, faster rollout timelines, and lower resistance during deployment.
Assessing Your Level
For each statement below, score honestly: Yes, Partially, or No.
- We assess organisational readiness before deploying AI capabilities
- We have a defined change management methodology for AI initiatives
- We track adoption metrics (usage, satisfaction, task completion) for every AI deployment
- Champions are identified and supported in each team that receives AI
- Communication about AI changes is targeted by stakeholder group
- Leaders at all levels understand their role in supporting AI adoption
- We learn from each deployment and improve our change approach
- Change management runs in parallel with technical delivery, not after it
- Adoption outcomes are connected to leadership performance metrics
- Our change approach adapts based on team-level data
Scoring:
- 0-3 Yes → Level 1
- 4-5 Yes → Level 2
- 6-7 Yes → Level 3
- 8-9 Yes → Level 4
- 10 Yes → Level 5
Technology maturity gets all the attention. Change maturity determines whether anyone uses the technology you build. Assess both. Invest in both. The organisations that build AI change capability alongside AI technical capability will outperform the ones that treat adoption as an afterthought.