Most Coaches Are Wrong About AI

57.7% of coaches say AI has no place in coaching. They're not wrong about the current state of AI coaching. They're wrong about what comes next. The research points to something surprising: the gap between coaches who embrace AI and those who reject it isn't about mindset. It's about literacy.

That number comes from the EMCC and Henley Business School's 2024 study. Nearly six in ten coaches looked at AI chatbots delivering coaching and said: no. Not real coaching. Not what we do.

They're correct. A chatbot that asks "How does that make you feel?" after every response is not coaching. A large language model that generates goal-setting frameworks without context, without relationship, without the ability to sit with silence - that is not coaching either.

But here's the problem with being right about the wrong question. The question is not "Can AI replace coaching?" The question is "What happens when AI enters the coaching relationship whether you invited it or not?"

Because it already has. 48% of coaching clients are using ChatGPT between sessions. Not because they prefer a chatbot. Because their coach isn't there from Monday to Friday, and something is.

What does the data actually say?

Let's start with the numbers, because they tell a story that the headlines miss.

The EMCC/Henley Business School survey (2024) found that 57.7% of coaches don't believe AI chatbots can deliver proper coaching. The ICF Global Coaching Study (2025) found that 37% of coaches view AI as a direct threat to the profession. And here's the demographic detail that matters: 88% of ICF-credentialed coaches are aged 45 or older.

That's not an insult. It's a data point. And it matters because the next finding explains why.

Diller, Czarniawska, and colleagues (2024) studied how coaches respond to AI at a neuropsychological level. What they found was striking: AI triggers what researchers call "behavioural inhibition" in coaches - a measurable pattern of higher perceived threat and lower curiosity compared to other professional groups.

Think about what that means. Coaches - people who spend their entire careers helping others navigate change, uncertainty, and discomfort - respond to AI with the exact pattern they help their clients overcome. Higher threat. Lower curiosity. Avoidance.

"The irony is not lost on the researchers. Coaches are experts in helping others navigate precisely the kind of response they themselves are having to AI." - Diller et al., 2024

And here's the worst part: it's self-reinforcing. When you avoid something because it feels threatening, you never learn enough about it to reduce the threat. The avoidance confirms the fear. The fear drives more avoidance. Behavioural inhibition becomes behavioural paralysis.

Is the gap really about mindset?

This is where it gets interesting. Because the obvious conclusion - "coaches need to change their mindset about AI" - is wrong.

Haase and colleagues at the Weizenbaum Institute (2025) studied AI adoption across professions and found something that should change how we talk about this entire topic. The single strongest predictor of whether someone adopts AI is not their attitude toward technology. It's not their age. It's not their openness to change. It's their understanding of how AI actually works.

AI literacy. That's it.

People who understand what a large language model does - and crucially, what it doesn't do - are significantly more likely to see it as a tool rather than a threat. People who don't understand it project their fears onto it. The fear isn't irrational. It's uninformed. And uninformed is fixable.

This reframes the entire conversation. The coaching profession doesn't have a mindset problem. It has a literacy problem. And literacy problems have a solution that mindset problems don't: education.

Haven't we seen this before?

Yes. Almost exactly.

Between 2015 and 2020, financial advisors faced robo-advisors. Betterment, Wealthfront, and a wave of automated portfolio management tools arrived, and the advisory profession split into two camps.

Camp one: "My clients want a human relationship. They'll never trust an algorithm with their money." Camp two: "We're all going to be replaced. The robots are coming."

Both were wrong.

What actually happened was more nuanced. Robo-advisors were excellent at commodity tasks - portfolio rebalancing, tax-loss harvesting, basic asset allocation. They were terrible at everything that required judgement, context, and relationship - estate planning, navigating divorce, helping a founder decide when to sell.

The advisors who thrived were the ones who understood the technology well enough to use it. They automated the routine work. They spent more time on the conversations that mattered. Their clients got better service. Their practices grew. They didn't adopt robo-advisory because they loved technology. They adopted it because they understood what it could and couldn't do.

The advisors who refused to engage lost clients. Not to robots. To other advisors who offered both the human relationship and the technological efficiency.

The coaching profession is on the same curve, roughly five years behind. And the pattern is identical: the differentiator isn't enthusiasm for AI. It's literacy about AI.

What does behavioural inhibition actually look like?

Let's make this concrete, because it shows up in coaching conversations every day.

A coach hears about an AI tool for session notes. Their first thought is not "How does that work?" Their first thought is "What if it replaces me?" That thought triggers avoidance. They don't look into it. They don't try it. They tell themselves they're protecting the coaching relationship. What they're actually doing is protecting themselves from uncertainty.

This isn't weakness. This is neuroscience. The Diller et al. (2024) research shows that AI activates threat-detection systems in coaches more strongly than in other professional groups. The researchers believe this is precisely because coaches understand the relational nature of their work - they know what's at stake.

But understanding what's at stake is different from understanding what's actually happening. And that gap - between emotional response and informed assessment - is the literacy gap.

A coach who understands how a large language model works knows three things:

• It cannot form a relationship. It generates text based on patterns. It does not care about your client. It has no stake in the outcome. This means it will never replace the coaching relationship. It literally cannot.
• It is excellent at pattern recognition in text. Themes across ten sessions. Language patterns that suggest a shift in confidence. Recurring goals that haven't been acted on. A coach doing this manually takes hours. The AI does it in seconds. That's not threatening. That's useful.
• It requires supervision to be safe. Left unsupervised, an AI will generate plausible-sounding content that may be wrong, inappropriate, or out of context. With a coach reviewing every output, it becomes a tool. Without supervision, it's a liability.

A coach who knows these three things responds to AI very differently from a coach who doesn't. Not because their mindset changed. Because their understanding did.

Why does the coaching profession have an advantage here?

Here's what's strange about this situation. Coaches are arguably better equipped to navigate this transition than any other profession. And yet they're among the slowest to engage with it.

Think about what coaches do every day. They help clients examine assumptions. They help clients distinguish between a feeling and a fact. They help clients move from avoidance to engagement. They help clients develop new capabilities in the face of uncertainty.

That is exactly the process required to close the AI literacy gap.

The coach who says "AI threatens the coaching relationship" is making a statement that sounds like a fact but is actually a feeling. A well-trained coach would recognise this in a client. The question is whether they can recognise it in themselves.

And the advantage goes further. Coaches who understand AI - who close the literacy gap - can help their clients navigate AI disruption in their own organisations. The executive who's anxious about AI replacing their team. The leader who doesn't know how to set AI strategy. These are coaching conversations that are happening right now, and the coach who understands AI can hold them with credibility.

Avoiding AI doesn't just limit your practice tools. It limits your relevance to clients who are dealing with AI every day.

What does closing the gap actually require?

Not much, frankly. That's the good news. The research from Haase (2025) is clear: understanding how AI works is the predictor. Not becoming a data scientist. Not learning to code. Understanding.

Specifically, three things close the gap:

1. Understanding what AI is and isn't. A large language model predicts the next word in a sequence. That's it. It's very good at this, which is why its outputs sound intelligent. But it doesn't think, feel, or understand. Once a coach grasps this, the threat level drops because the reality is less scary than the projection.
2. Seeing the supervision model in action. Most coaches who resist AI are reacting to the replacement model - chatbots coaching clients without a human in the loop. Once they see a supervision model, where the coach reviews and approves everything, the conversation changes. The tool starts to look like what it is: an assistant, not a replacement.
3. Trying it with one client. The Diller et al. (2024) research shows that behavioural inhibition decreases with exposure. The avoidance cycle breaks when you actually engage. Not with every client. Not with your most complex case. With one client, for one engagement, with full control over every output.

That's education first, tools second. Understand it, then try it. In that order.

"I was sceptical until I saw that I approve everything before it reaches the client. That changed the whole conversation for me. It's not AI coaching. It's my coaching, with AI helping me do more of it." - Executive coach, Amsterdam

What happens if the gap doesn't close?

The same thing that happened to financial advisors who didn't engage. They didn't get replaced by robots. They got outperformed by peers who used robots as tools.

A coach without AI support spends 30 minutes writing session notes from memory. A coach with supervised AI has accurate notes in five minutes. The first coach is not a worse coach. But they have less time for the work that matters.

A coach without AI support loses the thread between monthly sessions. A coach with supervised AI sends coach-approved reflection prompts, accountability nudges, and continuity notes between sessions. The first coach is not less caring. But the client experiences less continuity.

A coach without AI support spots patterns through intuition and experience. A coach with supervised AI has intuition, experience, and a system that surfaces language patterns, theme progression, and developmental arcs across an entire engagement. The first coach is not less skilled. But they have less information.

Multiply this across a practice. Across a year. Across a career. The gap compounds.

And the clients notice. Not because they want AI in their coaching. Because they want the outcomes that AI-supported coaching delivers: faster notes, better preparation, more continuity, deeper pattern recognition. They want a coach who shows up to every session having already reviewed the themes, the progress, the sticking points. They don't care how it happens. They care that it happens.