How to Use AI for Coaching Practice: My Experience with Claude AI

The Professional Challenge

As an LMSW who integrates food allergy counseling with ICF coaching, when appropriate, I work on both behavioral and performance goals with clients. This dual approach requires clear transitions between modalities for both me and for my clients, and I was curious about using AI to sharpen this skill, especially as this will be part of our presentation this fall at the Columbia University Coaching conference.

Last week's Coaching.com talk by Jonathan Reitz introduced transcript analysis for professional development. I decided to expand on this concept by practicing with Claude AI using a real session transcript (all names scrubbed for confidentiality). It was an experiment that yielded valuable insights about my ongoing coaching skill development.

What the Analysis Revealed

I provided Claude AI with an anonymized transcript from a food allergy counseling session where I'd integrated coaching elements. The analysis was useful:

Strengths identified: Strong counseling presence, appropriate trauma-informed approach, building on client's existing resources.

Coaching development opportunities: Evoke more client discovery and more powerful coaching questions.

For example, when my client said "I don't know how to convince my brain this wasn't about me" and mentioned “hypnosis,” I immediately suggested guided meditation and other hypnosis-based techniques. While I did ask about her past successful strategies later in our session, the coaching opportunity immediately would have been to get curious about her existing resources first, before offering any solutions.

The Practice Experiment

I decided to role play with Claude AI, with the AI taking my client's role while I practiced pure coaching skills.

Initial attempt: Claude AI delivered perfect insights with unrealistic clarity and immediate breakthroughs, which I noted to the AI is nothing like actual human clients.

I redirected Claude AI: "Make this more realistic. Real clients are confused, genuinely stuck, and less articulate."

The adjusted practice created authentic coaching challenges for me the coach: staying curious when someone truly doesn't know, resisting the urge to provide solutions, trusting the client's capacity for discovery.

A Key Breakthrough Moment

During the more realistic practice with Claude AI, when my “client" expressed feeling "permanently stuck" about dating challenges, I noticed my impulse to reassure and explain.

Instead, I asked: "If that were true, then what?"

This question helped the “client” discover her own vision of possibility—feeling safe with someone, dating from hope rather than fear. The insight emerged from her process, not my guidance.

Professional Takeaways

For coaches using AI practice tools:

Demand realistic difficulty. Perfect AI responses don't build skills. Request genuine confusion, resistance, and human messiness from your AI and keep pushing it.

Focus on your process. The goal isn't successful AI sessions—it's practicing specific skills like staying curious, asking powerful questions, and trusting client wisdom.

Notice the distinction. Coaching empowers clients to access their own answers rather than receiving expert guidance.

For potential coaching clients:

Your wisdom matters. Effective coaching assumes you have more answers than you realize.

Confusion is productive. Good coaches stay with your uncertainty rather than rushing to clarity.

Ownership creates change. Insights you discover stick better than advice you receive.

The Result: Pause & Get Curious

This Claude AI practice was a great reminder about how to transition between counseling and coaching modes and now I have a sticky note at my desk that reads: "Pause and get curious."

It's a simple shift that honors both my professional training and my clients' inherent wisdom.

Interested in exploring what you already know? Let's discover what's possible together.

Next
Next

Presenting at the 5th Bi-Annual Columbia Coaching Conference