Reflection and professional judgement

Working with an AI reflective partner

An AI reflective partner is a tool you work with to examine your coaching. It might bring a recurring intervention to your attention, offer another interpretation of a moment, or help you explore a question about your practice.

The appeal is having something to think with between conversations with peers or a supervisor. For an independent coach, that can be valuable.

But the usefulness depends on what kind of coaching it assumes you are trying to do. An assessment that rewards quickly reaching an action could miss the value of staying with uncertainty. Feedback can sound authoritative while applying a standard you would not choose for that conversation.

That makes developing a reflective partner a matter of judgement as well as setup. Building and refining one is a practical part of AI-Augmented Coaching.

When convincing AI feedback gets your coaching wrong

Useful feedback has a defensible connection to what happened. Its tone and fluency are not enough to establish that connection.

During our course testing, a coach described receiving useful feedback alongside an interpretation she knew was wrong: the model read a client’s direct language as a sign of upset. She had been in the conversation and knew the person. The words alone did not carry everything she knew.

That is a useful way to approach AI feedback. An observation can be worth exploring without its interpretation being correct.

The difficulty is that useful observations and mistaken interpretations can arrive in the same response. Accepting or dismissing the whole answer loses that distinction. Your knowledge of the relationship remains part of the judgement.

There is another trap here: criticism can feel more credible simply because it is uncomfortable. A model that has stopped flattering you has not necessarily become a better judge of your coaching.

I would treat the response as material for reflection, with room to disagree. The value lies in what becomes clearer about your work, including when you decide the model has misunderstood it.

When reflection starts to get in the way of coaching

Useful reflection helps you notice something, understand a choice or decide what to explore next. It also needs to fit within the attention and time you have.

One participant described using a reflective partner after every session and then noticing its influence during coaching: she was critically analysing herself rather than remaining in the flow of the conversation. She subsequently used it more selectively.

Other course discussions have raised a related problem. AI can generate additional observations and avenues of inquiry until a short reflection becomes a substantial piece of unpaid work.

I would pay attention to both effects. Are you becoming clearer about your practice? Are you able to return to the client with greater attention? Or are you accumulating interpretations and monitoring yourself against them?

Being able to generate more reflection creates a new demand on the coach: deciding what deserves attention. A tool can be impressive while the overall way of using it becomes unsustainable.

There is no universal frequency I would prescribe from these experiences. The judgement is whether reflection is serving your development and the relationship, at a cost in time and attention you can sustain.

Providing context for reflection

Enough to make sense of the question. More information does not automatically produce a more useful reflection.

Your account of the client shapes the response. If you describe someone as avoiding responsibility, an answer organised around avoidance may feel like independent confirmation. It may partly reflect your framing.

Notice what you are putting into the conversation as well as what the model gives back. Ask what context the question actually requires before adding more client information.

Bringing AI reflection to human supervision

An AI reflective partner may help you prepare something to take to supervision: a recurring intervention, a question about your judgement or a moment you want to understand more fully.

It cannot provide the relationship or professional accountability of a human supervisor. A text-based review also misses aspects of what happened in the room.

You can take the AI’s interpretation—and your response to it—to supervision. Why did it feel convincing, uncomfortable or easy to dismiss? That reaction may be useful material in its own right.

(Before using client material, see AI, confidentiality and your clients’ trust.)