I use AI. Every day. For writing, for research, for stress-testing my own thinking. I’m not going to pretend it isn’t useful – it is. If you want to brainstorm career options, draft a difficult email, or work through a logic problem, AI is genuinely good at that.
So why would you pay a human coach when you could type the same question into a chatbot for free?
Because coaching isn’t about the question. It’s about everything around the question.
In the late 1960s, psychologist Albert Mehrabian conducted research into how people communicate feelings and attitudes. His findings – often cited, usually badly – suggested that when we’re communicating how we feel about something, only 7% of the message is carried by the words. 38% is in the tone of voice. 55% is in facial expression.
Those numbers have been widely misapplied. They don’t mean that all communication is 93% non-verbal. Mehrabian’s research was specifically about emotional communication – how we convey feelings and attitudes. He’s spent decades trying to correct the misquotation.
But here’s what matters for coaching: coaching is emotional communication. When you’re working through whether to leave a job, how to handle a manager who’s blocking your progression, or why you keep saying yes to things that make you miserable – you’re not exchanging information. You’re processing how you feel about your situation. And in that context, Mehrabian’s research is directly relevant.
A human coach hears the 93%. The pause before you answer. The slight shift in your voice when you mention a particular colleague. The way you sit forward when you talk about the work and lean back when you talk about the organisation. The thing you keep circling back to without ever quite saying directly. The throwaway comment at the end of the session that turns out to be the most important thing you’ve said all hour.
AI gets the 7%. The words you type.
That’s not a small difference. That’s the difference between someone helping you with what you said, and someone helping you with what you meant.
I’ve lost count of the number of times a client has come to me with a specific problem, only for us to discover that the real issue is something they couldn’t see from the inside – a pattern they’d fallen into, an assumption about their situation that had never been tested, or a systemic problem that isn’t theirs to fix alone and needs other people in the organisation to act.
There’s something else, too. Being heard – properly heard, by another human being who is paying full attention to you and has no agenda other than helping you think – is not a minor thing. For a lot of my clients, it’s the first time in their professional lives that someone has listened without simultaneously working out how to manage them, redirect them, or explain to them why their feelings aren’t quite right.
One of my clients put it simply: knowing he’d spoken to a real person about what he was going to do meant he couldn’t just avoid it afterwards. AI doesn’t create that kind of accountability. You can close the tab. You can’t close the conversation you had with someone who actually heard you.
AI is a tool. A good one for certain situations. But coaching is a relationship – and the relationship is the intervention. No algorithm does that. I’m not sure one ever will.
If you’re weighing up whether coaching is worth it, or whether you could get the same thing from a chatbot, try this: think about the last time someone asked you how you were, and you said “fine” – and they knew you weren’t. That’s what a human coach does. That’s the 93%.
If you want the numbers, download The Business Case for Coaching.
