AI Model Wiki

    Skip to main content
    15 min
    Lesson 6 of 6

    Common Pitfalls & How to Avoid Them

    From the course AI for Coaches & Mentors

    The mistakes coaches make when adopting AI — and how to avoid undermining your practice.

    Pitfall 1: Using AI to Avoid the Hard Work of Coaching

    What happens: A coach uses AI to generate all session questions, send all follow-ups, and produce all reflections — and gradually loses touch with the client as a real person. The coaching becomes efficient but hollow. Clients feel it and don't renew.

    Why it happens: AI makes it easy to produce output. Quantity of output is not the same as quality of impact.

    How to avoid it:

    • Use AI for the administrative and mechanical: summaries, templates, scheduling, resource lists
    • Never use AI to generate the core insight or reframe — that comes from listening, not prompting
    • A test: could you describe your client's current biggest fear and what they need to hear right now, without looking at any notes? If not, you may be delegating too much to AI
    • Keep at least one touchpoint per week that is entirely unscripted — a voice note, a brief check-in call, a handwritten quote

    Pitfall 2: Confidentiality Breaches Through Careless AI Use

    What happens: A coach pastes a client's full name, employer, relationship issues, and career fears into ChatGPT to generate a reflection prompt. This data is now potentially in a training dataset.

    Why it happens: The convenience of general-purpose AI tools makes it tempting to share rich context.

    How to avoid it:

    • Anonymize client data before using any AI tool that isn't enterprise-grade (remove name, employer, identifiable details)
    • Use Claude for Enterprise or ChatGPT Enterprise if you're using real client context — these have contractual data privacy protections
    • Include an AI data usage policy in your coaching agreement: explain what AI tools you use and how you protect client data
    • Never paste raw session transcripts into consumer AI tools without anonymization

    Pitfall 3: Sending Unreviewed AI Output to Clients

    What happens: A coach auto-sends an AI-generated session summary without reading it. The summary contains an error — it misattributed a quote to the client that the coach actually said — and the client feels misrepresented.

    Why it happens: The efficiency gains of AI tempt people to remove the human review step.

    How to avoid it:

    • Treat every AI output as a first draft that requires human sign-off before sending
    • Build review into your workflow, not optional: "AI generates → I review → I send" is a non-negotiable sequence
    • For anything emotionally sensitive (progress reviews, difficult feedback), write it yourself with AI as a reference only

    Pitfall 4: Tool Overload in the First Month

    What happens: An enthusiastic coach signs up for Otter.ai, Notion AI, Mem.ai, ChatGPT Plus, Claude Pro, and Jasper in the same week, spends $200/month, and uses none of them consistently.

    How to avoid it:

    • Start with one tool, use it daily for two weeks, evaluate honestly
    • Add a second tool only after the first is habit
    • Your initial stack should be: one transcription tool (Otter.ai) and one AI writing assistant (Claude or ChatGPT) — that covers 80% of coaching AI use cases
    • Consolidate before expanding: you rarely need more than 3-4 AI tools in a coaching practice

    Key Takeaways

    • •AI handles the mechanical; the core insight and reframe must come from you — that is what clients pay for
    • •Anonymize client data before using consumer AI tools; use enterprise tools for sensitive context
    • •Every AI output needs human review before reaching a client — never skip this step
    • •Start with one tool used consistently; add a second only after the first is embedded habit
    0 of 6 lessons marked complete

    Progress is kept in this browser only.