You have never had more access to good advice than you do right now.
Every framework, every technique, every bit of wisdom about how to live is a sentence away – free, instant, and endlessly available in a search engine or LLM like ChatGPT.
And yet people are no less stuck than they were.
AI can do a lot of what gets sold as coaching…
It’s worth being honest about how much. Whatever AI can’t quite manage today, it will manage soon, and a great deal of what people call coaching is already well within reach.
Setting goals and breaking them into steps, maintaining weekly accountability, providing a tidy summary of what you said last time, naming the obvious limiting beliefs that run in patterns, and reframing them can all be done now with ChatGPT and a handful of purpose-built coaching apps.
For plenty of people, it works.
I’m not offended by any of it. I’ve never been much use at the cookie-cutter version of the work anyway: the fixed method, the technique you run people through, the performance coaching that’s really a process with a person attached. If your coaching is essentially a process, the process is the part that automates, and I’d hand most of it over without a fight.
So if you’re asking whether AI will replace life coaching, here’s the honest answer. Most of it, yes – the broad, teachable middle of it, and sooner than coaches would like.
But the part people actually came for was never in that middle.
Coaching was never about information.
We already have the information – all of it, more than any human who ever lived, free and instant. And still we are no less lost. Which points at something the industry would rather you didn’t dwell on: the information was never the thing doing the work.
You’re not stuck because nobody has told you the four steps to get unstuck. You can find those in thirty seconds, and you probably already have. A coaching conversation does something else. It’s where you write the story you tell yourself about your own life, and, sometimes, where you rewrite it.
Therapy can help you understand that story; but coaching is where you take hold of it and write what comes next.
There’s a good book called Personality Isn’t Permanent, and the title is the whole point. The narrative you carry about who you are isn’t fixed. It’s flexible. You can change it. And when it’s tied to your values and sense of purpose, changing it can change a great deal.
So the real question isn’t whether AI can run a session, it’s this:
Do you want a machine to help write your life, or a human being – with judgment, with nuance, with their own years of living behind them – to help you shape it?
The question a machine can’t ask
There are two kinds of questions.
One kind is looking for an answer. What’s the capital of France? What should I eat? How do I structure my week? AI is extraordinary at these. At its core, it’s an answering machine: you give it a question, and it answers it. That’s the design.
The other kind isn’t looking for an answer at all. Who am I? What is this for? What would I do if I weren’t afraid? These aren’t problems to be solved. They’re questions you sit inside, and the sitting is the point. They open a space rather than close one. In coaching, as in Buddhism, the best questions work this way: not as retrieval, but as a catalyst for creativity.
A system built to resolve questions is the wrong tool for the questions that aren’t meant to be resolved. It will try to help. It will hand you an answer. And the answer is the very thing getting in your way.
You can’t see your own eyes
This is the part I’d want you to take seriously, even if you skim the rest.
You can see with your eyes, but you can’t see your eyes.
Every one of us carries an unconscious, constructed self, built by family, school, culture, and mostly by economic interests that were never especially concerned with our flourishing. It filters everything. The Sufis called it the commanding self, and you can’t see it, because you’re seeing through it.
A great deal of coaching is putting pressure on exactly that: investigating the conditioning, finding its edges, rewriting it through your own insight.
Now think about what an AI model actually is. It’s trained on the consensus, a vast and fluent average of everything we’ve already said and already think. Which makes it, near enough, made of the commanding self.
This is the line between the reframing I mentioned earlier and the real thing. A model can flip the obvious beliefs, the ones that run in patterns, all day long.
What it can’t reach is the deeper layer, the conditioning you can’t see, because that layer is exactly what it’s built from. It can reflect your conditioning back at you in lovely sentences. It can’t stand outside it, because it has no outside. To see the water you’re swimming in, you need someone who isn’t the water. You can’t bolt that on. It’s structural.
“You’re not broken” is a stance, not a setting
There’s a premise underneath the way I work. You’re not broken; you’re obstructed. Something whole in you already knows what it wants, and the job isn’t to fix you. It’s to clear what’s in the way so your own health can express itself. That orientation comes out of the human potential movement, and it’s something you decide to hold, not a default you fall into.
It matters here because a machine optimizing toward a goal will, left to itself, reach for the opposite posture: find the problem, close the gap, fix the thing. That’s the stance of most of the internet it learned from. You can instruct it to behave otherwise, and it will perform the lines well. But holding the belief that the person in front of you is already whole, and letting that belief shape every question you ask, isn’t a setting you can toggle. It’s a way of seeing someone, and I’m not sure it survives being faked.
What the future actually looks like
Now the bigger picture, because I owe you my real view of it and not a comfortable one.
AI isn’t arriving by accident. It’s arriving because the systems we’ve built to run the world have grown too complex for us to manage on our own.
In that role, it’s useful, and I don’t say that grudgingly. Used well, it can lift an enormous amount of mechanical and administrative work, and the hopeful version of this is real: more of us get our time back, with room to make things, to think, to be with the people we love, to do the work that was actually ours to do.
But a tool that’s brilliant at running systems can, at the very same time, be beside the point for the meaning of a human life. Those are different questions. The danger isn’t that AI is bad. The danger is quieter: that we let the thing that’s good at the systems start answering questions it was never built for – what’s worth wanting, how to live, who to become – and don’t notice we’ve handed them over.
If anything, the noise is about to get louder. Frames will shatter faster. More people will find themselves in the fog, that gap between one chapter and the next, where the old map is gone, and the new one hasn’t formed yet. The fog is exactly where this work matters, and the one place the answering machine has nothing to offer, because the fog isn’t a question with an answer. It’s a passage you have to walk.
The way through hasn’t really changed. It’s slow. It happens in language, with another person, in the middle of an actual life. It doesn’t scale, and it was never supposed to.
The honest version
So here’s my honest answer about AI and the future of this work.
For most people, and most of what they want from coaching, the machine will be enough. I mean that without bitterness. If it helps more people think a little more clearly about their lives, good.
But if you’re after something larger, to make your life whole, to line up your values and your work and your days and the story you tell about all of it, that was never an information retrieval problem.
No system can hand it to you, however convincing it may sound, because getting there and doing the work are one and the same act.
