Every knowledge worker
AI-skilled
House in order. People AI-skilled.
Everyone has AI.
Almost no one uses it well.
Every organization has AI licenses by now. Most employees barely use them, or use them badly. Not because they don't want to, but because nobody taught them how.
Researchers at Harvard and BCG gave consultants AI tools for their work. On tasks where AI fell short, consultants using AI performed worse than colleagues without it. The output looked convincing. It just wasn't right. Nobody caught it.
You see this pattern everywhere. Doctors who used an AI tool for three months became measurably less accurate. Not because the tool was bad, but because they stopped thinking for themselves.
Then there's the problem nobody talks about: junior staff working with AI deliver faster output. But the experience that would've made them great at their craft? They never build it. The output is there. The expertise isn't.
AI without skills doesn't make people more productive.
It makes them more dependent.
If you want AI to deliver real value in your organization, you need two things.
Get your house in order.
AI works so well in software development because everything is documented. Every line of code, every change, every repository. The knowledge is structured and accessible, so AI can get straight to work.
Knowledge work is different. Expertise sits trapped in PDFs, in people's heads, in email threads, in folders nobody opens anymore. Without that structure, every AI tool delivers mediocre results, no matter how advanced the model.
Get your people AI-skilled.
People who know how to write a clear prompt. Who give AI enough context to come back with something useful. Who spot when the output is wrong. And who fold AI into how they actually work, because they've seen it help.
Everyone's talking about new models and the AI hype. Meanwhile, most organizations are stuck on the basics. Employees get a Copilot license and use it for summaries when they could be transforming their entire workflow.
The problem isn't the technology. The house isn't in order, and the people aren't AI-skilled.
What we choose
We choose
Building skills over buying licenses
Changing behavior over transferring knowledge
Real work over demos
Knowing where AI fails over blind trust
Human agency over automation
Taking responsibility over producing output