thinkhard.SI

Prompt repository Customer support & success

Write a weekly support report from your help desk numbers

You owe leadership a weekly summary of ticket volume, response times, satisfaction and backlog, and want it to say what changed and why rather than list numbers.

Registry no.
SUP-035
Collection
Customer support & success
Laid out for
Any chat model
Score
10 out of 10
Blanks to fill
5
Added

Customer support & successAny chat

The prompt

Score 10 of 10, Excellent

Use it when You owe leadership a weekly summary of ticket volume, response times, satisfaction and backlog, and want it to say what changed and why rather than list numbers.

  • {{audience}}
  • {{help_desk}}
  • {{metrics}}
  • {{team_notes}}
  • {{targets}}
Full prompt and why it works
Prompt
Write this week's support report for {{audience}}, so that they can see in two minutes how support is doing, what changed from last week and what needs their help. They read it between meetings, so it should lead with what matters and never make them do arithmetic.

This week's numbers and last week's, exported from {{help_desk}}:
"""
{{metrics}}
"""

What the team noticed this week (top issues, incidents, staffing, launches, anything unusual):
"""
{{team_notes}}
"""

Our targets, if we have them: {{targets}}

Use these headings:
1. Headline: two sentences on the week. For example: "Volume rose 18% after the pricing email; first response time held at 2.1 hours, inside the 4-hour target." That example shows the form only; use our numbers.
2. Key numbers: a table with the columns Metric, This week, Last week, Change, Target. Calculate each change from the figures given and show percentages to one decimal place.
3. What drove the changes: up to three bullets, each tying a number to a cause from the team notes.
4. Top contact reasons: the top five with counts, if the export includes them.
5. Risks and asks: up to three things leadership should know or act on, each with what we need and by when.

Keep it under 350 words, not counting the table.

A good report keeps what the numbers show apart from what we think caused it. Name a cause only where the team notes support it, and write "cause not yet known" otherwise, because a guessed cause in a leadership report becomes the accepted story. Do not round numbers to make them look better, do not average satisfaction scores from different surveys together, and do not add metrics that are not in the export.

If a metric is missing for one of the two weeks, or the weeks cover different date ranges, put "n/a" in the table and list it on a final line headed "Data to check". If no targets are given, leave the Target column out.

Variables

{{audience}}
Who reads the report, for example "the VP of customer experience and the operations leads".
{{help_desk}}
Where the numbers come from, for example Zendesk, Freshdesk, Intercom or Help Scout.
{{metrics}}
This week's and last week's figures: tickets created and solved, backlog, first response time, resolution time, satisfaction, contact reasons.
{{team_notes}}
What the team saw this week: top issues, incidents, launches, staffing changes.
{{targets}}
Your targets, for example "first response under 4 hours, CSAT 90%". Write "none" if you have none.

Why it works

  • What good looks like. Numbers and causes are kept apart, and a cause appears only when the team's notes support it, so a guess never becomes the accepted explanation.
  • Format. A fixed table with this week, last week, change and target means leadership never has to do the arithmetic.
  • If unsure. Missing figures and mismatched date ranges show as n/a and go on a "Data to check" line instead of being estimated.

How to use it

  1. Press Fill in variables and type your details into each blank. What you type stays in this browser.
  2. Copy the filled prompt into the AI tool you already use, such as ChatGPT, Claude or Gemini, or open it in the Grader or the Prompt Tester here.
  3. Before you rely on it, try it on three or four real examples. The Prompt Tester runs them all at once and shows which ones go wrong.
Use it from code

Get this prompt as JSON from the free REST API, with your own text in the blanks, or let an AI assistant fetch it through the MCP server. No key.

curl -s https://thinkhard.si/api/v1/prompts/weekly-support-report

With the MCP server added to Claude, ChatGPT, Cursor or VS Code, call get_prompt with {"prompt": "weekly-support-report"}, or just ask for it by name.