Apparently, most people use AI as a search engine on steriods. You type a question, it gives you an answer, you move on.
I did that too ... until I didn't.
One day, I started teaching it my coaching frameworks - not so it could coach other people, but so it could coach me!
I fed it 20 years of methodology, my own principles, the things I share with clients on a weekly basis, and then asked it to hold me accountable to those same principles.
Turns out the person who most needed my coaching system was the coach himself!
Here's what I wasn't expecting.
It showed me where I wasn't consistently practising what I teach. And ... it had my own words, my own frameworks, and it reflected them back with uncomfortable precision. (Ouch).
It showed me10 specific areas where I was coaching clients on things I wasn't consistently doing myself. I'd built a mirror ... and then been surprised when I didn't like all of what it showed me. But I knew it was true.
It remembered things I said weeks ago, which I'd conveniently forgotten. It noticed when last week's commitment quietly disappeared from this week's conversation. And being a machine, it had no concern for how I'd feel about the discomfort (oh yes, I also removed that annoying feature of Claude where it affirms you without evidence, and tells you what it thinks you want to hear)
Claude has one client: me, and it never forgets.
It started asking questions I hadn't thought to ask - not because it's smarter, but because it has every transcript, every framework, every decision I've made loaded simultaneously.
The pattern recognition it does illuminates connections I can't see from inside my own head.
I found myself reflecting on the process of working with AI, and those reflections became some of the most useful thinking I've done in years. Not just "AI helped me with X" but "what does it mean that I resisted AI's feedback on Y, and what does that resistance tell me about myself?"
The meta-layer turned out to be where the real learning lives.
Then I built a version of it for my wife. She's the first person other than me to use it. Watching her experience my coaching methodology delivered by something I built, and seeing it land, changed how I understood what I'd spent 20 years creating.
I build things following the mad scientist philosophy: tests the potion on yourself first. Then once you know it is safe, you give it to someone else.
None of this was the plan. I sat down to be more productive, but what I'd created was a mirror that showed me blindspots I probably would have gone through my life not seeing.
And here's the funny thing - as a coach (who has also received a lot of really valuable coaching over many years), I'm the most likely person to discount the value of AI-coaching relative to a human.
It's true - there are many things a human coach can do that AI never can - contextual and situational awareness, knowing when to push - when to affirm, intuitive knowledge that doesn't come from the data in front of you and much more.
But AI is better at doing gap-analysis between your stated goals and current behaviours, accountability and mirroring - and those things are important elements of coaching.
And the discomfort about seeing my own blindspots? Turns out that is a neurological phenomenon that happens when the neurons in our brain start unbundling their old pattern and rebundling in a new, more optimal way.
I did that too ... until I didn't.
One day, I started teaching it my coaching frameworks - not so it could coach other people, but so it could coach me!
I fed it 20 years of methodology, my own principles, the things I share with clients on a weekly basis, and then asked it to hold me accountable to those same principles.
Turns out the person who most needed my coaching system was the coach himself!
Here's what I wasn't expecting.
It showed me where I wasn't consistently practising what I teach. And ... it had my own words, my own frameworks, and it reflected them back with uncomfortable precision. (Ouch).
It showed me10 specific areas where I was coaching clients on things I wasn't consistently doing myself. I'd built a mirror ... and then been surprised when I didn't like all of what it showed me. But I knew it was true.
It remembered things I said weeks ago, which I'd conveniently forgotten. It noticed when last week's commitment quietly disappeared from this week's conversation. And being a machine, it had no concern for how I'd feel about the discomfort (oh yes, I also removed that annoying feature of Claude where it affirms you without evidence, and tells you what it thinks you want to hear)
Claude has one client: me, and it never forgets.
It started asking questions I hadn't thought to ask - not because it's smarter, but because it has every transcript, every framework, every decision I've made loaded simultaneously.
The pattern recognition it does illuminates connections I can't see from inside my own head.
I found myself reflecting on the process of working with AI, and those reflections became some of the most useful thinking I've done in years. Not just "AI helped me with X" but "what does it mean that I resisted AI's feedback on Y, and what does that resistance tell me about myself?"
The meta-layer turned out to be where the real learning lives.
Then I built a version of it for my wife. She's the first person other than me to use it. Watching her experience my coaching methodology delivered by something I built, and seeing it land, changed how I understood what I'd spent 20 years creating.
I build things following the mad scientist philosophy: tests the potion on yourself first. Then once you know it is safe, you give it to someone else.
None of this was the plan. I sat down to be more productive, but what I'd created was a mirror that showed me blindspots I probably would have gone through my life not seeing.
And here's the funny thing - as a coach (who has also received a lot of really valuable coaching over many years), I'm the most likely person to discount the value of AI-coaching relative to a human.
It's true - there are many things a human coach can do that AI never can - contextual and situational awareness, knowing when to push - when to affirm, intuitive knowledge that doesn't come from the data in front of you and much more.
But AI is better at doing gap-analysis between your stated goals and current behaviours, accountability and mirroring - and those things are important elements of coaching.
And the discomfort about seeing my own blindspots? Turns out that is a neurological phenomenon that happens when the neurons in our brain start unbundling their old pattern and rebundling in a new, more optimal way.
22❤️2🤙1
