Most AI upskilling programmes fail because they're owned by IT. IT understands the technology but not the workflows, the people, or the change dynamics. HR understands all three. The organisations closing the AI skills gap fastest are the ones where HR took ownership — not as a training function, but as a change management function.
What AI literacy actually means at enterprise scale
'AI literacy' is often treated as binary: people either know how to use AI or they don't. In practice it's a three-tier capability. Awareness means understanding what AI can and can't do in your specific function. Application means using AI tools effectively in day-to-day workflows. Advanced means building or adapting AI workflows and evaluating new tools for the team. Most enterprise programmes are designed for awareness training when the real competitive advantage lies in application.
How to prioritise which roles to upskill first
Not every role has equal leverage in an AI-enabled world. The roles with the highest return on AI upskilling share three characteristics: high-volume repetitive tasks, access to structured data, and output that needs to be adapted for multiple audiences. In most organisations, this points to finance, marketing, HR operations, and sales — ahead of leadership, which should be second-wave.
The three mistakes most training programmes make
- Generic content that doesn't connect to actual role workflows — people complete it and then don't change how they work
- One-time events instead of embedded habits — skills learned in workshops decay within 30 days without reinforcement
- Measuring completion instead of application — programme success metrics that incentivise ticking a box rather than changing behaviour
Building for the second wave
The first wave of AI literacy — helping employees understand what AI is — is largely over. The second wave is harder: helping employees fundamentally change how they work. That requires HR to go beyond training design into job architecture, performance frameworks, and team structure. The organisations thinking about this now will have a structural advantage in two years.
The companies that win on AI aren't the ones with the most tools. They're the ones where the most people know how to use them well. That's an HR problem, not an IT one.