Tool literacy is not improvement capability
Using AI tools effectively is different from knowing what organisational work is worth improving and whether change made it better.
There is an understandable enthusiasm for AI training. Teach people how to use ChatGPT or Copilot. Show them how to prompt effectively. Demonstrate summarisation, drafting, analysis and automation. People quickly find things they can do. That is useful. But it does not mean the organisation has learned how to improve. I have seen staff become perfectly capable of generating marketing copy, writing emails and producing documents with AI while the underlying work remains unchanged. Sometimes the results are merely generic. Sometimes they create new problems. A polished document can still contain an inappropriate commitment.
A well-written customer response can still be wrong. A faster activity can still be unnecessary. This is because tool literacy and improvement capability are different things. Tool literacy helps somebody obtain useful outputs from AI. Improvement capability helps an organisation decide what needs changing, why it needs changing and whether the intervention actually made things better. The distinction matters because AI is an unusually flexible technology. Once people know how to use it, they can find potential applications almost everywhere. Without an improvement method, experimentation can become the strategy. I think there is a more interesting possibility.
AI itself may help organisations develop some of the improvement capability they lack. It can interrogate process descriptions. It can structure conversations with employees. It can compare different accounts of work. It can identify assumptions. It can ask what evidence is missing. It can help design small investigations. People still provide organisational knowledge, gather evidence, exercise judgement and decide what should change. AI scaffolds the investigation. This may be particularly valuable for smaller organisations without Lean specialists, process-improvement teams or business analysts. So perhaps the AI skills question needs broadening. We should certainly teach people how to use AI. But we should also teach organisations how to use AI to ask better questions about themselves. Because knowing how to operate the tool is only the beginning. The more valuable capability may be knowing what is worth improving in the first place.