Most coaches end a call, feel like it went well, and move on. But “felt like it went well” is a guess, and guesses are how good clients quietly drift away. The honest signal you need is sitting in your client’s head right after the session, and it fades fast.
This coaching feedback survey prompt builds a short, well-ordered post-session survey in a couple of minutes. You give the AI your niche, the session type, and what you most want to learn, and it returns a survey built the way good surveys are actually built: easy questions first, the right mix of scales and open text, and explicit permission for the client to be critical. By the end of this page you’ll also understand why it pulls honest answers instead of polite ones.
When to use this
- You finish sessions thinking they went well but have no real data to confirm it.
- You want a repeatable feedback loop after every call, not a once-a-year survey.
- You’re refining a new offer or session format and need fast signal.
- You want testimonials and referrals but feel awkward asking directly.
- You’re trying to spot at-risk clients before they ghost or cancel.
The prompt
Copy this whole block into ChatGPT, Claude, or Gemini:
You are an expert in coaching client experience and survey design. Your job is to build a short post-session feedback survey that gets honest, useful answers without annoying my client.
Before writing, ask me up to 3 clarifying questions if anything below is unclear. Otherwise, proceed.
CONTEXT
- My niche: {{NICHE}}
- The session type clients just finished: {{SESSION_TYPE}}
- What I most want to learn from this feedback: {{GOAL}}
- Where the survey will live: {{TOOL}}
- My tone: {{TONE}}
TASK
Design ONE feedback survey that:
1. Opens with one warm sentence explaining why their honest answer helps them get more out of coaching.
2. Contains 5-7 questions total, ordered easy-to-deeper.
3. Includes at least one 1-5 rating scale, one multiple-choice question, and one open-text question.
4. Has at least one question that gives the client permission to be critical (e.g. 'What is one thing I could do better next time?').
5. Ends with a single optional open box for anything else.
6. Labels each question's type so I can drop it straight into {{TOOL}}.
CONSTRAINTS
- Takes under 2 minutes to complete.
- Plain, human language. No jargon, no leading questions that fish for praise.
- Matches my tone: {{TONE}}.
- Do not invent client names, stats, or testimonials.
After the survey, give me 2 short reminder messages I could send to nudge clients who haven't responded.
How to customize it
Replace the five {{VARIABLES}} before you send it:
| Variable | What to put | Example |
|---|---|---|
{{NICHE}} |
Your specific coaching niche | career coaching for mid-career women |
{{SESSION_TYPE}} |
The kind of session clients just had | 60-minute 1:1 video coaching call |
{{GOAL}} |
The one thing you most want to learn | whether clients leave with a clear next step and feel heard |
{{TOOL}} |
Where the survey will live | Google Forms |
{{TONE}} |
How you sound | warm, casual, low-pressure |
See it in action (full example)
Here is the exact prompt, filled in for a career coach. This is the whole input, nothing hidden:
You are an expert in coaching client experience and survey design. Your job is to build a short post-session feedback survey that gets honest, useful answers without annoying my client.
Before writing, ask me up to 3 clarifying questions if anything below is unclear. Otherwise, proceed.
CONTEXT
- My niche: career coaching for mid-career women
- The session type clients just finished: 60-minute 1:1 video coaching call
- What I most want to learn from this feedback: whether clients leave with a clear next step and feel heard
- Where the survey will live: Google Forms
- My tone: warm, casual, low-pressure
TASK
Design ONE feedback survey... [rest of the prompt unchanged]
And here is the complete output the model returns:
Quick clarifying check (1 question): Do you want this survey to be anonymous, or tied to the client’s name so you can follow up personally? I’ll assume named, with an option to skip their name unless you tell me otherwise.
Survey title: How did today’s session land?
Intro line: Thanks for showing up today. Two minutes of honest feedback helps me tailor our next session to exactly what you need, so please be straight with me, the good and the awkward both help.
1. (1-5 rating) Overall, how valuable was today’s session for you? 1 = not really useful · 5 = exactly what I needed
2. (1-5 rating) How clear are you on your next step before we meet again? 1 = still foggy · 5 = crystal clear, I know what to do
3. (Multiple choice) Which best describes how you felt during the call? – Heard and understood – Mostly heard, a little rushed – I had more to say than we had time for – I didn’t feel like I got to the real issue
4. (Open text) What was the most useful thing you’re taking away today?
5. (Open text) What is one thing I could do better next time? (No wrong answers here, this is genuinely how I improve.)
6. (Multiple choice) How are you feeling about your progress overall right now? – On track and motivated – Moving, but slower than I’d like – Stuck and could use a check-in – Not sure
7. (Optional open text) Anything else on your mind? Leave it blank if not.
2 reminder messages to nudge non-responders:
“Hi [Name], no rush at all, but if you have 2 minutes the feedback form from our last call is here: [link]. Even one line helps me make our next session better for you.”
“Hi [Name], reopening this gently, your honest take on our last session would genuinely shape the next one. Here’s the quick form whenever you have a moment: [link]. And if now’s not the time, just ignore me.”
That is ready to paste into Google Forms. Each question is labeled with its type, the order moves from easy ratings into honest open text, and question 5 explicitly invites criticism so you get real answers, not flattery.
Why this works
Three LLM principles are doing the heavy lifting here. Learn them and every prompt you write gets better:
- Role priming. The opening line (“You are an expert in coaching client experience and survey design”) tells the model which slice of its knowledge to draw from. “Make me a survey” pulls from generic form templates; naming a survey-design expert pulls from question-ordering, scale design, and bias-avoidance. Always assign a role that matches the job.
- Specificity in, specificity out. The model can only be as specific as your input. “Get me feedback” yields bland questions. Telling it your exact goal (“whether clients leave with a clear next step and feel heard”) makes the AI write questions 2 and 3 that measure precisely that. The sharper your
{{GOAL}}, the sharper the survey. - Constraints are quality control. Each rule removes a known failure mode. “No leading questions that fish for praise” stops the classic mistake of writing “How amazing was your session?”, which produces useless, flattering data. “Under 2 minutes” protects your response rate. And the “ask me up to 3 clarifying questions first” line lets the model fill gaps by asking, such as whether the survey is anonymous, instead of guessing wrong. That single line is the biggest upgrade you can make to any prompt.
Do this now
- Copy the prompt above into ChatGPT or Claude.
- Replace the five variables with your real niche, session type, goal, tool, and tone.
- Send it, and answer any clarifying question it asks.
- Paste the questions into your form tool and send it after your very next session. Real data beats a good feeling.
Pro tips
- Keep it to one survey per session type. A discovery call and a 12th coaching session need different questions. Run the prompt once for each.
- Always keep question 5. The “what could I do better” question is where retention insight lives. Cutting it for comfort defeats the purpose.
- Review answers weekly, not yearly. Patterns across three or four responses tell you more than any single rave review.
- Turn a strong open-text answer into a testimonial ask. When someone writes something glowing, that’s your cue to follow up and ask permission to quote them.
0 comments
No comments yet.