# Understand a client report — your turn

Try this after watching the lesson. You manage a client's training account, and your manager wants a recommendation before the next session is scheduled.

## The client's email

**Subject: Pilot training — what should change?**

Hi,

The attendance figures look encouraging: 90 people registered and 72 attended the pilot training. But team leaders are still getting questions about how to use the process, so I don't think attendance tells us whether we're ready to roll this out.

We received 24 feedback forms. Eighteen of those respondents said they couldn't finish the practice task during the session. Several people asked for a worked example they could refer to afterward. I can't tell whether we tried to cover too much, the exercise needed more explanation, or people simply needed more time. We haven't checked which it was.

Before we book another session, can you recommend what to change? We'd like people to leave able to do the task themselves. Please include how we could check whether the next session works better, rather than just counting who turns up.

Thanks,
Taylor

## Your task

1. Ask ChatGPT to explain the email in plain English and suggest a recommendation for your manager.
2. Improve that same answer. Separate the client's requests, your recommendations, and what still needs checking. Ask for one immediate action and one longer-term improvement, each supported by the email.
3. Compare the answer with the original email. Make one direct edit if the recommendation needs it. Use the check guide below to review your reasoning.

Use this practice text for the exercise. If you later work with a real client email, follow your employer's rules for sharing information with AI tools.

## Check guide

- **What matters:** Attendance is not evidence that participants can perform the task independently. The client wants practical learning improvements and a way to evaluate them.
- **Keep the counts straight:** 18 of the 24 feedback respondents reported not finishing the practice task. That is 75% of respondents, not 75% of all 72 attendees. The report does not establish what the other attendees experienced or whether the other six respondents finished.
- **Immediate recommendation:** Review the incomplete practice work or ask a few participants and the trainer where they got stuck. This helps distinguish between the possible explanations the client lists. A worked example is also worth considering because participants explicitly requested one; it should support practice, not replace investigating the problem.
- **Longer-term recommendation:** Use what that review reveals to adjust a small pilot session, then check whether participants can complete the task themselves. Compare task performance as well as attendance. Neither an improved slide deck nor a higher attendance rate by itself establishes better learning.
- **Keep uncertainty visible:** The cause of the incomplete work is unconfirmed. The email does not specify available trainer time, budget, a target pass rate, or an agreed date for another session. Propose a practical approach without inventing commitments.
- **There is no single required answer:** A different recommendation can work if it addresses an explicit concern, uses the evidence accurately, and distinguishes known facts from proposals.

## Keep learning

[Return to AI for Real Work](https://www.youtube.com/@LearnAIForRealWork) for another lesson.

Ready to turn your recommendation into a clear, effective email? [Watch the email lesson](https://youtu.be/YyMd5bnTuOA).
