Managers are under a peculiar kind of pressure. They are told AI will transform their work and that they must lead the change. Many are privately unsure where to start.
Underneath sits a harder question. Am I capable of leading when the tools change faster than I can learn them?
Why managers matter so much
Gallup's 2026 State of the Global Workplace report found manager-led adoption was a top driver of frequent AI use. In US data, 79% used AI frequently when their manager actively supported it. Otherwise the figure was 46%.
Yet fewer than a third of employees say their manager actively supports AI use. Manager engagement also fell from 31% to 22% in three years.
Microsoft's 2026 Work Trend Index found that where managers modelled AI use, trust in AI agents rose by 30 points. Confident managers create confident teams.
Two kinds of worry
Some worry is about competence: can I use these tools well? Training, practice and peer support help here.
Other worry is about relevance: will my role still matter? Answering that takes experience of what good human work with AI looks like.
The framework below works on both. The early stages build competence. The last stage lets managers answer the relevance question by example.
Build this capability in your organisation
Growth Performance designs bespoke programmes that turn insight into measurable behaviour change.
Four stages: anxiety, curiosity, confidence, capability
Stage one: name the anxiety. Make space for managers to say what worries them. Unspoken fear drives hesitation and avoidance. Saying "I do not understand this yet" is where learning starts.
Stage two: build curiosity. Give managers hands-on time with the tools their teams use, in low-stakes settings. Add peer learning, so they hear what worked and what failed.
Stage three: build confidence. Confidence comes from small wins. Help each manager use AI for one real decision, then reflect on it with a peer.
Stage four: develop capability. Capability means leading others through the same journey. A manager who has done the work can model learning and normalise mistakes.
Make it real
Give managers protected development time as part of the role. Psychological safety matters at this level too. Reich and colleagues found in 2026 that safety predicted whether employees started using AI at all.
Judge success by whether each team makes AI part of its real work. Speed of rollout is a poor guide. That is the Adoption condition.
Try this
Use one AI tool every day for a week. Note where it helps and where it disappoints.
Then share what you learned with your team, including your mistakes. Take the Adoption Condition Diagnostic to see where to focus next.
