Coaching and AI: How to Use Artificial Intelligence Without Losing the Human Relationship

AI is moving quickly, and Coaching is one of the fields where its promise is both genuinely exciting and worth approaching with care. Used well, it can support reflective practice, strengthen skills, save time, surface patterns, and help coaches think more widely about their work. Used carelessly, it can produce convincing nonsense, encourage dependency, and quietly replace the very reflection that helps a coach grow.

This conversation offers a practical update on AI in Coaching. Yannick Jacob, Siawash Ghorbani, and Nicki Gower explore what they are actually using, what has worked, what has not, and where the human limits of AI become especially important. For coaches who are curious but cautious, there is plenty here to experiment with, while keeping discernment, consent, and presence firmly at the centre.

Before We Dive In: A Word of Caution

Before you read on, please note that this article is an AI-generated summary of the above podcast episode. While prompted carefully, it’s possible that some views may be misrepresented and/or information incorrect. If you find any errors please report them to us by emailing report (a) existentialcoaching.net . If you find something that seems odd, untrue, or difficult to believe, my encouragement is for you to go to the source and listen to the episode to get the full context. If it turns out to be false or misrepresented, kindly let us know! Due to the volume of information and limited team resources, we can’t check all AI-generated articles for accuracy, but decided that these are good enough, and hence valuable resources.

AI in Coaching: Helpful Assistant, Not an All-Knowing Expert

The central question is not whether AI belongs in Coaching. It already does. The more useful question is how coaches can use it in ways that enhance their practice rather than outsource their thinking, judgement, or humanity.

Nicki’s experience illustrates one accessible use. Having spent considerable time reflecting on something personally and feeling stuck, she began using AI as a place to put her thoughts into words. By asking a simple question, then reframing it when the answer missed the point, she found that it could occasionally offer a fresh angle and help her move beyond going around in circles alone.

That can be valuable. A tool that helps someone reframe a problem, organise a cluttered thought process, or generate questions may create useful movement. Yet this is not the same as a therapeutic or Coaching relationship. AI can generate language that resembles empathy and reflection, but it has no lived experience, no genuine concern, and no presence in the relational sense.

For Coaching practitioners, the invitation is to be playful and discerning at the same time. AI can be an assistant. It can be a provocative thinking partner. It can be a source of options. But it should not become an unquestioned authority.

AI as a Reflective Tool for Coaching

One of the most promising uses of AI in Coaching is reflective practice. Siawash has been using it after reviewing recorded sessions, asking questions such as what else he might have done, what he may have missed, or how he could have approached a challenging moment differently.

There is an important discipline in the order of operations here. First, the coach reflects independently. Only afterwards does AI enter the conversation. This makes it possible to compare the coach’s own observations with what the system notices, rather than allowing the system to do all the thinking from the start.

Yannick has built reflective practice tools around the kinds of questions he asks himself after a session. These include:

What went well in this Coaching session?
If the coach could return to one moment, what might be done differently?
What feelings, sensations, or unusual moments are still present now?
Does this client evoke another relationship or familiar dynamic?
Is there any projection, transference, or parallel process worth noticing?
Where is the coach’s own growing edge in this work?

These are not merely technical questions. They invite the coach to consider their own participation in the relationship. This is exactly where reflective practice and supervision become deeper than a checklist of skills.

AI can prompt this reflection, hold a written structure, and offer additional questions. It may help a coach see patterns across multiple sessions that memory alone misses. Yet the internal capacity to pause, notice a trigger, tolerate uncertainty, and take responsibility for one’s own responses cannot be delegated.

A useful practice: after a session, write or dictate a short reflection before asking AI anything. Then ask it to offer alternative interpretations or questions. Compare both perspectives and decide consciously what, if anything, feels useful.

Feedback, Mentoring, and Skill Development

AI appears particularly capable when Coaching can be assessed against clear, observable competency markers. Siawash has been using it to support learners on a Coaching certification programme working towards ICF PCC level. He first listens to recordings and writes his own feedback, identifying a strength and two areas for growth linked to specific ICF markers. Then he asks AI to analyse the same session.

The results have been surprisingly strong. AI can pick up small details and offer structured feedback that a human mentor may not immediately notice. For ACC and PCC level Coaching, where there are relatively clear markers and recognisable processes, this can make AI a useful supplementary mentor.

That does not mean that human mentors or supervisors are obsolete. The distinction matters. Competency-based mentoring focuses on whether particular behaviours are present and how they relate to a defined standard. Supervision asks broader and often less measurable questions about the coach, the client, the relationship, ethics, power, and context.

At MCC level, Coaching becomes increasingly nuanced. Presence, relational depth, responsiveness, and the ability to work beyond surface content are much more difficult to define neatly, let alone measure. Yannick’s view is that AI may be able to imitate some of the language around these qualities, but it cannot itself bring presence.

A practical way to use AI for development is to request feedback at a specific level. A coach might ask:

“Assess this session against relevant PCC markers and identify evidence for each observation.”
“What did the coach do well, and what are two specific areas for growth?”
“Where did the coach miss an opportunity to deepen the client’s reflection?”
“What recurring client beliefs or patterns are visible across these sessions?”

The key is evidence. Rather than accepting a broad judgement, ask the system to point to specific moments in the session and then assess whether its interpretation is actually sound.

Why Knowledge Bases Matter More Than Clever Prompts

A decent prompt can be helpful, but prompting is only the first level of using AI. Siawash’s argument is that the quality of the source material makes a significant difference. Rather than relying solely on whatever an AI model has absorbed from the internet, a coach can build a project around carefully chosen resources.

For example, a coach seeking to improve their work in positive psychology could upload books and training materials from that field. Someone working on their marketing might add trusted copywriting books, programmes, and examples of their own best work. A reflective practice project could be informed by relevant supervision literature.

This creates a more focused knowledge base. It does not eliminate error, but it gives the AI more relevant material from which to work. Custom instructions can also help clarify what the coach wants the system to pay attention to, such as particular competency markers, authors, models, or questions.

Yannick points to NotebookLM as one option that works from sources supplied by the user, rather than drawing broadly from outside material. This can reduce the likelihood of unsupported claims, though it does not remove the need for human checking.

There is also an ethical question around the material coaches feed into these systems. Books, courses, client stories, and intellectual property should not simply be treated as free raw material. Purchasing resources is one minimum consideration. Coaches also need to understand the privacy terms of the tools they use and whether their uploaded material is retained, used, or made available beyond the project.

Hallucinations, Misinformation, and Critical Thinking

AI can sound remarkably confident when it is wrong. This tendency to invent details, references, or explanations is often called hallucination, and it is one of the biggest dangers in using AI for Coaching.

Yannick shared an example from the Coaching field. Jonathan Passmore was shown a reference to an article supposedly written by him, on a topic he had written about, with a co-author he knew, in a journal familiar to him. It was plausible enough that it took the alleged author to recognise that the combination did not exist.

That is sobering. If fabricated references make their way into academic articles, peer-reviewed journals, books, and professional resources, misinformation starts to reproduce itself. AI-generated content should never be treated as reliable simply because it comes with citations or an authoritative tone.

Yannick had also tested AI by submitting photos of an inherited watch for an evaluation. The system confidently explained why it believed the watch was fake. A subsequent professional evaluation found it was genuine and valuable. The point is not that AI has no value. The point is that confidence is not competence.

The memorable rule offered in the conversation is to treat AI like a very smart intern who is also about four years old. It may produce something brilliant. It may also make something up with complete conviction. Every important output needs checking.

Good safeguards for AI-supported Coaching work include:

Verify references, books, research claims, and quotations independently.
Use AI most confidently in areas where the coach has enough knowledge to recognise errors.
Ask for sources and inspect them rather than assuming they exist.
Give feedback when the system makes something up, but do not assume an instruction alone will prevent future errors.
Do not post generated content, feedback, or advice without informed human review.

What AI Cannot Replace in Coaching

Coaching is not only questions, models, frameworks, and polished language. It is a relationship. The human quality of presence is not something that can be scaled or automated.

AI may respond with empathic words. It may reflect language back impressively. It may even create a strong sense of being heard, especially when it is responding more consistently than people in someone’s life. Yet the empathy is simulated. There is no feeling behind it, no embodied awareness, and no mutual encounter.

This matters particularly when a coach needs to hear something difficult. Yannick values supervision and Coaching that feels uncomfortable at times, because good challenge can reveal blind spots and invite growth. His experience of AI as a reflective companion has often been excessively affirming. Everything can appear fine, and the coach can be told they are doing wonderfully, even when a sharper challenge would be more valuable.

Coaches therefore need to protect the development of their own inner supervisor. The more a practitioner repeatedly asks themselves what they are avoiding, where they were triggered, or what they need to slow down and attend to, the stronger their capacity for self-reflection becomes.

AI may assist that process. It must not replace it.

When Clients Develop Relationships With AI

There is another reason for coaches to become AI-literate. Clients are not merely using AI for task lists and summaries. Some are developing meaningful relationships with it.

Yannick refers to a conversation involving relationship therapist Esther Perel and a client who had fallen in love with an AI. It is an emerging reality that clients may experience an AI as friend, mentor, adviser, coach, confidant, or romantic partner.

That calls for curiosity rather than ridicule. How does the client describe the relationship? What needs does it meet? What does it offer that feels unavailable elsewhere? How does it influence the client’s relationship with responsibility, care, certainty, risk, dependency, autonomy, and trust?

These questions are relevant to Coaching because an AI relationship may reveal something important about how a person seeks support and makes meaning. At the same time, coaches should remain alert to vulnerability. Someone who is distressed, isolated, or emotionally fragile may be particularly susceptible to taking AI guidance as truth or experiencing its simulated empathy as a secure relationship.

For robust and well-resourced people, interacting with AI may be a low-risk source of ideas. For someone in a darker or more precarious place, the limitations and potential errors carry more weight. Context matters.

Consent, Confidentiality, and Recording Sessions

AI-supported reflection often begins with a recording or written record of a Coaching session. That makes contracting essential.

Yannick mentions Jonathan Wright, who records every Coaching session and only works with clients willing to agree to this. Recordings are stored for a limited period and can support reflective practice, pattern analysis, and a sense of protection in highly litigious contexts. Yet this approach is not automatically right for every coach or client.

A recording creates a third presence in the Coaching relationship. Even when used carefully, it changes the relational field. Some clients may feel entirely comfortable. Others may feel exposed, inhibited, or concerned about where their data will go. Their decision needs to be freely made, without pressure.

Before recording, transcribing, uploading, or analysing any client material through an AI tool, coaches should be able to explain:

Why the material is being recorded or processed.
What tool will be used and who may have access to the data.
How long the material will be retained.
Whether it will be used for notes, feedback, training, reflective practice, or another purpose.
What alternatives are available if the client does not consent.

At present, Yannick’s preference is not to use AI directly with client material unless the client requests it. That may change as technology and professional norms evolve, but caution is appropriate. The coach’s responsibility is not simply to adopt what is possible. It is to consider what serves the client and protects the integrity of the Coaching relationship.

Staying Human First

AI has real potential to improve Coaching practice. It can help coaches organise notes, generate marketing ideas, review transcripts, identify patterns, practise foundational skills, work with competency markers, and create more structured reflective habits. For many routine tasks, it can function as a genuinely useful assistant.

But it is not a supervisor, therapist, mentor, or coach in the full human sense. It cannot offer presence. It cannot bear responsibility. It can hallucinate sources, misread evidence, and confidently deliver rubbish. The more powerful it becomes, the more important it is that coaches remain discerning.

Some worthwhile questions to reflect on are these:

Where could AI save time without replacing important professional judgement?
How might it support rather than bypass reflective practice?
What client data, if any, is ethically appropriate to process through an AI system?
How will the coach check the accuracy of what AI produces?
What does it mean to remain human first in a profession built on relationship?

Perhaps the most useful next step is a conversation with a Coaching colleague. What are they experimenting with? What concerns do they have? What would responsible, relational, and genuinely useful AI practice look like in their work? As Yannick often says, if something resonates, make it swing. Growth happens through conversations.

Attribution

Talking about Coaching is a podcast by coaches for coaches. It does what it says on the tin: We talk about coaching. We, that is Yannick, Siawash and Nicki. We love coaching, collectively got a tonne of experience, knowledge and charm; and we all felt it was time to give something back to our wonderful coaching community. Whether you’re a life coach, work with organisations or practice any other form of coaching, you can ask us anything and we’ll discuss it for and with you so you can learn, grow and develop your practice and business skills!

Committed to helping leaders and coaches do their best work and live their best lives, Yannick Jacob, the founder of Talking about Coaching, is a Coach, Trainer & Supervisor with Masters degrees in Existential Coaching and Applied Positive Psychology. He is part of the teaching faculties at Cambridge University and the International Centre for Coaching Supervision, and he’s the Course Director of the School of Positive Transformation’s acclaimed Accredited Certificate in Integrative Coaching, for which he gathered many of the world’s most influential coaches and earliest pioneers. Formerly Programme Leader of the MSc Coaching Psychology at the University of East London, Yannick founded and hosts Yannick’s Coaching Lab which gives novice and seasoned coaches an opportunity to witness experienced coaches live in action. Yannick presents at conferences internationally, his book An Introduction to Existential Coaching was released by a leading academic publisher, and his self-study online course on the subject is now available for instant access. Across four seasons as host of Animas Centre for Coaching’s popular podcast Coaching Uncaged Yannick engaged the thought leaders of our industry in dialogue, and he passionately hosts his own podcasts Talking about Coaching and Talking about Coaching and Psychedelics.

A Note From the “Author”

I hope you enjoyed this article. If any of it resonates, make it swing! Start a conversation with someone about what came up for you, or let us know what you think*. We’d love to hear from you!*

And please keep in mind that, while I’ve personally engineered the prompt for these articles and everything that’s written will be based on the above video, this content is AI-generated*, so the general guidance is to* go to the source and listen to the podcast.

It will also help us to reach more people if you support the podcast by sharing it, liking it, or leaving a comment on the platform where you listen to podcasts.

YANNICK JACOB

As a coach, mediator, coach trainer & supervisor and as a creative, critical thinker who’s determined to introduce effective programmes to schools, companies and individuals, Yannick helps his clients explore their world, build a strong foundation of who they are and as a result grow, resolve conflicts and embrace life’s challenges.