The Gap: What I Tell Leaders About AI vs How I Was Actually Using It

In the keynotes, I talk about the AI adoption gap. The distance between where most organisations think they are with AI and where they actually are. The difference between having tools installed and genuinely changing how you think.

I've delivered this message probably a thousand times. I believe it completely. And for the better part of two years, I was living on the wrong side of it.

The stage version

On stage, I describe three stages of AI maturity in organisations. At the first stage, AI is a productivity tool; you use it to go faster on things you were already doing. At the second stage, AI starts to reshape workflows, and how it changes, not just the speed. At the third stage, AI becomes strategic infrastructure; it changes the questions you can ask and the decisions you can make.

I tell leaders that most of their organisations are stuck at stage one, even though they believe they're at stage two. That the jump from tool to infrastructure requires intentionality, not just access.

Then I fly home, open my laptop, and ask AI to draft a follow-up email.

The mirror version

The AI audit I ran on myself, the one I described in post two of this series, made this undeniable. I was operating at stage one. Comfortably, efficiently, professionally. But stage one.

I was using AI to produce faster. Not to think differently. Not to ask better questions. Not to challenge my own assumptions or expand what I was capable of seeing.

This is, I've come to understand, an extremely common pattern for people who speak or write about AI professionally. You develop a fluency in describing the transformation. The description becomes the thing. And the actual practice, the messy, uncertain, sometimes uncomfortable work of genuinely integrating AI into how you think, doesn't keep pace.

What I changed

The shift I made was deliberate and uncomfortable. I started treating every significant decision, problem, or piece of work as an AI collaboration, not a solo effort with AI assistance at the end. I'd start with AI. Challenge it. Build on what it gave me. Argue with it. Push it harder.

The keynotes got sharper. The frameworks got tested differently. The advice I gave to clients became more honest about where I'd personally experienced the friction, not just where I'd observed it.

I'm not claiming I've solved this. But I've closed the gap enough that standing on stage doesn't feel like a performance anymore. It feels like a report from the field.

Practical AI: Audit your prompts

Go back through your AI conversations from the last week. Read your prompts. Not the outputs, the prompts themselves. Ask:

• Am I asking AI to do things, or to think with me?

• Do my prompts reflect what I actually believe about AI's potential, or are they much more basic than that?

• Where is the gap between what I tell others to do with AI and how I'm actually using it?

  
  

The prompts are a window into your actual relationship with AI, not the one you'd describe if someone asked.

An interesting exercise is to paste a week's worth of prompts into a document and then upload it as context into a different tool. Ask it to challenge you, to find new perspectives, and show you ways to see your own train of thought differently.