Building a culture that embraces experimentation, encourages responsible innovation, overcomes siloed decision-making, creates psychological safety, and balances speed with governance is the least visible — and most decisive — factor in whether AI adoption actually sticks.
Where this gets hard
- Building a culture that embraces experimentation is difficult in organisations historically rewarded for certainty and consistency.
- Encouraging responsible innovation requires holding two values — speed and caution — in tension, which is uncomfortable for most teams.
- Siloed decision-making prevents AI learnings and use cases from spreading across the organisation.
- Psychological safety is essential for employees to admit when an AI tool isn't working, but is rarely explicitly cultivated.
- Balancing speed with governance is treated as a zero-sum trade-off rather than a design challenge.
Where to start
- Reward well-reasoned experiments that fail, not just experiments that succeed, to build genuine appetite for trying new things.
- Create a small, visible forum for sharing AI use cases and lessons across business units.
- Explicitly signal psychological safety by having leaders share their own AI mistakes or misjudgements.
- Design lightweight, fast governance for low-risk AI use, reserving heavier governance for genuinely high-risk use cases.
- Measure and reward responsible AI use, not just usage volume or speed of adoption.
The consulting document includes a culture health checklist and a set of workshop prompts for surfacing hidden resistance or silos.
When was the last time someone in your organisation was publicly praised for an AI experiment that didn't work?