The biggest risk with AI right now isn’t technical. It’s cultural. Deloitte’s public stumble last week, after an AI-generated report error hit headlines, says it all. I was quoted in the Australian Financial Review, “AI did it” isn’t a defence (see the article here). Comment here.
Staff are confused and uncertain
In most companies, generative AI is already in use. A finance analyst automates commentary. Marketing uses prompts to generate first drafts of creative copy. But it’s often happening quietly. People aren’t sure what’s allowed.
Leaders I talk with often assume AI adoption is progressing. Yet when I ask, few can say who is using what, where the risks are, or whether usage aligns with policy. The gap between visibility and reality is widening.
Global data confirms this. Only 30% of employees say their organisation has policy or guidance for using generative AI at work (Melbourne Business School & KPMG global survey 2025).
This creates uncertainty for staff in a world that is confusing, where the skills required to do their jobs are shifting. Massive, fast shifts. The World Economic Forum’s 2025 report says employers expect 39% of workers’ core skills to change by 2030. And whichever survey I look at, the vast majority want AI training.
The implications of AI uncertainty and eroding skills
Culture is a pattern of behaviors: the more aligned they are with your strategy, the greater your chance of success. Uncertainty and eroding skills create fear. This is a key risk with AI. It’s not just about technology. While we can’t create certainty in an unpredictable world, leaders can be clear about the strategic direction.
If your strategy is to implement AI and digital capabilities, then it is simple: send clear behavioral signals to your team about AI.
Your culture is influenced by what leaders say and do about AI
3 Tips to Building AI-Enabled, Digitally Capable Culture
Here are some immediate small steps you can take to move your culture in the direction of AI:
1. Communicate AI expectations
Even if you haven’t finished a formal AI governance policy, the Executive team needs to communicate simple expectations about AI. Clear up confusion: Is experimentation allowed? Is it OK to use public tools? What corporate data can we share publicly, if any. Likewise, with AI, the cat is out of the bag, so guide appropriate usage with clear expectations.
2. Make AI practice part of the job
There are a variety of things organizations are doing to encourage using AI, such as holding an “AI week” or “lunch n learns” where people can share what they have learned. This reframes AI from risk to routine.
3. Share failures as well as success
Make it safe for people to share their misses, not just their wins with AI. When senior leaders share their own mistakes, it lowers the fear factor and signals support for experimentation.
These ideas build skills, shrink fear, increase clarity, and turn policy into behaviour. When people feel safe to try, fail, and share, curiosity replaces silence. And curiosity, done right, compounds into capability.
I teach AI Culture Intelligence inside The Turbochargers Hub in one-hour express modules (with recordings and editable playbooks). Email me if you’d like access or click here.
“Clear is kind. Unclear is unkind.”
Brené Brown
PS More Ways to Accelerate your Strategy
Looking to move your strategy forward? Lisa Carlin works with selected clients directly to go from strategy to plan to result. Email Lisa.
A bespoke in-house Bootcamp is the fastest way to learn. Get the skills your teams need to go from Strategy to Plan to Results. AI optional. Bootcamp info here.
Subscribe here to ensure you don’t miss this weekly newsletter.
Connect with Lisa on LinkedIn.
About Lisa Carlin
Lisa Carlin is a Strategy Execution Specialist. She works with business leaders to go from strategy to plan to result with culture intelligence and AI.
Lisa is author of the globally acclaimed newsletter, Turbocharge Weekly, read by 8,000 business leaders.
Lisa’s career includes roles at McKinsey and Accenture, then running her own business since 1999. Over this time she has delivered over 50 implementations with a 96% success rate.