Executive coaching is a harder case for AI than most coaching. Your clients are senior, their time is scarce, and what they tell you often includes things their own board has not heard yet. Get AI wrong here and the cost is not a clumsy answer. It is a breach of trust with someone who has very few people they can be candid with.
Get it right, though, and it gives back hours every week and makes the time between sessions far more useful to your clients. This guide covers where AI helps an executive coach, where it should stay out, how to keep confidentiality intact, and a handful of prompts you can use this week.
The short answer
AI helps executive coaches most outside the live session: preparing, reflecting, writing up, and supporting clients between sessions. It helps least, and carries the most risk, inside the conversation itself, where judgement, timing and trust do the work.
The rules that keep it safe are simple. Get consent before anything is recorded or transcribed. Know where client data goes and whether it is used to train AI models. Check the client organisation's own AI policy. Keep identifiable details out of general AI tools unless the terms protect them. And agree with the sponsor, up front, what any AI tool sees and what is shared.
Where AI helps an executive coach
Before a session
Preparing your thinking. Give an AI assistant your own anonymised notes from recent sessions and ask what patterns it notices and which questions might help the client see them. It will not know your client. It is a sparring partner for your preparation, not a source of answers.
Designing exercises. A new CEO who cannot let go of operational detail. A leader about to face a hostile board. AI is quick at drafting exercises, reflection questions and frameworks you can then adapt to the person in front of you.
Between sessions
This is where AI changes the most for senior clients. The moment an executive needs coaching is rarely the hour you have booked. It is the evening before the board meeting, or ten minutes before a difficult one-to-one.
An AI coach trained on your material can be there at that moment. A client writes "Board meeting tomorrow. How do I hold the room?" and gets an answer in your words, from your frameworks: open with the decision you need, then stop talking. Anything personal or sensitive goes back to you. There is more on how these work in AI clone for coaches.
Rehearsal. Executives can use AI to rehearse a hard conversation, with the AI playing a sceptical chair or a defensive direct report, then step out of the role and get feedback. Research at IMD with 167 executives found an AI tool built for this gave feedback on communication style that matched what human observers saw. The executives got the most from it when they asked specific questions and asked for examples.
After a session
Writing up. Turning your own notes into a short summary of what was agreed takes AI seconds. If you record or transcribe sessions to do this, that needs the client's clear consent, and you need to know where the recording goes.
Following up. A check-in on the action the client committed to, sent at the right moment, keeps momentum without you having to remember every commitment for every client.
In your own practice
Proposals and programme design. First drafts of proposals for sponsors, programme outlines and session plans.
Reflective practice. Describing a difficult moment from a session and asking AI to question you about it, the way a supervisor might. Useful between supervision sessions, never instead of them.
Where AI should stay out
The live conversation. The value of executive coaching is a skilled person paying full attention. A tool in the room, even a silent one, changes what a senior leader is willing to say. I would keep AI out of the session entirely.
Reading the politics. Who really holds power on the board, why the CFO is suddenly supportive, what the CEO is not saying about their successor. AI has none of that context and will fill the gap with confident, generic advice.
Judgement about when to challenge. Knowing when a senior leader is ready to hear something uncomfortable, and how to say it, is the core of the craft.
Assessing your client for the sponsor. If the organisation paying asks for an AI-generated view of how their executive is doing, say no. That is a fast way to lose the client's trust and the sponsor's.
Confidentiality: the executive coaching version
Executive coaching usually involves three parties: the coach, the executive and the organisation paying. Every AI tool you use adds a fourth, so it needs to be part of the agreement.
- Consent first. Nothing is recorded, transcribed or summarised by AI without the client knowing and agreeing.
- Know the data path. Where is the data stored, who can see it, how long is it kept, and is it used to train AI models? If you cannot answer, do not use the tool with client material.
- Check the client's policy. Many organisations restrict what their people can put into external AI tools. Your coaching should not quietly break that rule on their behalf.
- Keep identifiable details out of general AI tools. Anonymise names, companies and anything market sensitive before you ask for help with your thinking.
- Agree it with the sponsor. Put it in the coaching agreement: what any AI tool sees, what the sponsor receives, and what stays between you and the client.
The ICF's AI coaching standards cover much of this from the tool side. My plain English guide to the standards turns them into questions to ask any provider.
Which kind of AI tool for which job
Rather than a list of products, which goes out of date in months, here are the kinds of tools and what each is good for.
| Kind of tool | Good for | Watch out for |
|---|---|---|
| General AI assistant | Preparing your thinking, drafting exercises and proposals, reflective practice | Never paste identifiable client details unless the terms protect them |
| Meeting note-taker | Session summaries and action lists | Consent, where recordings go, and whether a senior client will speak freely |
| AI coach trained on your material | Supporting clients between sessions, in your voice, with personal topics passed back to you | It must answer from your material only and know when to hand over |
| Rehearsal tool | Clients practising hard conversations before they happen | It plays a role; it does not know the real people involved |
Prompts executive coaches can use this week
These work in any general AI assistant. Anonymise before you paste anything in.
Here are my anonymised notes from the last three sessions with a senior leader. What patterns do you notice? Suggest five open questions that might help them see those patterns for themselves. Do not give advice.
Design a 20 minute exercise for a newly promoted CEO who struggles to delegate to an experienced executive team. Use reflection questions, not instructions.
I coached a leader today who became defensive when I reflected back that they interrupt their team. Ask me questions, one at a time, the way a coaching supervisor would, to help me think about what happened.
Turn these notes into a one page coaching proposal for a sponsor. Include the aims, the shape of the engagement, how progress will be reviewed, and what stays confidential between coach and client.
And one for your clients to use before a hard conversation:
Play the chair of a board who is sceptical about my proposal to restructure the sales team. Push back the way a demanding chair would. After five exchanges, step out of the role and tell me where my case was weakest.
Questions executive coaches ask
What is the best AI for executive coaches? It depends on the job. A general assistant for your own preparation and writing, a note-taker only with consent, and an AI coach trained on your material for supporting clients between sessions. Judge each one on how it handles confidential data before anything else.
Should I tell my clients I use AI? Yes. Tell them what you use it for, what it never sees, and ask for consent where their information is involved. Senior clients respect a clear answer far more than a discovered surprise.
Can AI coach executives? It can support them, especially between sessions and for rehearsal. The research so far suggests people still rate human coaches higher where insight and the relationship matter. I have covered the studies in will AI replace coaches?
Is it safe to use AI with confidential business information? Only when you know exactly where the data goes, it is not used to train AI models, the client has agreed, and their organisation's policy allows it. When in doubt, anonymise or leave it out.
Where to start
Pick one job this week, outside the live session: preparing for your hardest client, or writing up your notes. Use AI for that alone, with anonymised material, and see what it gives you back. Then decide what is next.
If you'd like to see what an AI coach for your own practice looks like, I'll build a companion from your website, free. See your own companion
Robert, founder of Thinkable Systems