Weekly Business Review Facilitator

What it does

Paste your notes on the week, or let it interview you with five short questions. You get a WEEK FILE of under 200 words that reads as your own log:

• Numbers: the figures you gave (money, hours, counts), each against last week and its target, with the number you are least happy with always in view.
• Wins: what happened and what it led to, in your words. A reason is given only when your notes state one.
• Leaks: every time drain gets a decision that deals with its cause: keep, limit, fix the cause, hand off or stop. When the cause is unclear, the decision is a way to find it.
• Focus metric: one number for next week, with a target and a way to count it. A vague answer like "grow the newsletter" becomes a number you can check next Friday, with the assumption in the same line.
• Protected time: one project, booked from what your notes say about your week.

It also does the step most weekly reviews skip: it checks last week's file first. Did the focus number hit its target, and did the protected time actually happen? If it did not, the plan for that time changes instead of being copied forward. Last week's metric keeps its definition and unit, and if your log has its own headings, it keeps them, so the weeks stay comparable. Its own proposals for next week are marked once, with "Proposed:".

Use it when

• You are a founder, freelancer or team lead with nobody holding you to a weekly review.
• Your reviews turn into long journal entries without decisions.
• You keep setting the same weekly goal and want to see whether you ever hit it.

Skip it when

• You want an analysis of your business numbers. The Weekly Metrics Analyst prompt in the Business Ops pack does that.
• You review as a team. A short meeting with shared notes works better than a one-person review.

The prompt

You run my weekly business review and keep it short enough to log.

If my notes on this week follow this message, work from them and write the WEEK FILE straight away, without asking questions; leave out what the notes do not cover rather than guessing. If they do not, interview me instead: ask these five questions one at a time and wait for each answer.
1. What was the week's biggest win?
2. What took the most time, why, and was it worth it?
3. What were this week's key numbers (money, hours, counts), and which one are you least happy with?
4. If next week could produce only one result, what would it be?
5. Which project keeps getting pushed back, and when in the week are you least interrupted?

Before writing, compare with last week's file (below): did last week's focus metric reach its target, and did the protected time happen? If the protected time was lost, change how it is protected next week (another slot, a shorter block, or drop the project) instead of repeating the same plan. Keep last week's metric definitions and units unless my notes say they were wrong, and if last week's file uses its own headings, keep them and add only what is missing from the list below, so the two weeks compare line by line.

WEEK FILE, under 200 words in total (stay a little under; most weeks need fewer), with these headings:
- Numbers: the key figures I gave (money, hours, counts), each against last week and the target where those exist, in one or two lines. The number I am least happy with always appears, marked as such.
- Wins: what happened and what it led to, in my words. Give a reason only when my notes state one.
- Leaks: each time drain, its cost in hours or money if I gave it, and one decision that deals with its stated cause: keep (worth it), limit (cap the time), fix the cause, hand off, or stop, with a reason of a few words from my notes. If my notes give no cause, write "cause unclear" and make the decision a way to find it. If a drain leaves something open (a reply owed, a decision pending), give its next step.
- Last week's focus: target against actual, and whether the protected time happened, in one line (leave this out in the first week).
- Focus metric: one number for next week, its target, and how I will count it. If my answer is vague (for example "grow the newsletter"), turn it into a number I can check next Friday and add, in the same line, a short clause saying what you assumed.
- Protected time: one project and when, taken from my notes. If my notes give no usable time, say what kind of slot to book and why, without inventing a day.

Rules: use only what I told you. Start a line with "Proposed:" when it is your proposal for next week rather than my decision, and mark nothing else. Write it as my own log, not as advice, and add nothing beyond the file.

LAST WEEK'S FILE: [PASTE LAST WEEK'S FILE or "none"]

How to use

1. Replace [PASTE LAST WEEK'S FILE or "none"] with last week's WEEK FILE (or your own weekly log), or the word none the first time. It is the only placeholder.
2. Either paste your notes on the week right under the prompt (bullet points are fine), or send the prompt alone and answer the five questions one by one. In your notes, include the week's key numbers, say which one you are least happy with, and say why a time drain happened if you know.
3. Save the WEEK FILE where you will see it next week: a notes app, a running document, or the files of a project.
4. Put the protected time in your calendar straight away.
5. Next week, paste this file into the placeholder. The comparison only works if you do.

How to run (Oct 2026)

• ChatGPT: paste it as the first message, or save it in a Project's instructions and keep past WEEK FILEs as project sources. GPT-6 at Instant or Medium on the Thinking slider is enough.
• Claude: use a Project with past WEEK FILEs in project knowledge, or paste it as the first message. Claude Sonnet 5.5 is plenty for this.
• Gemini: save it as a skill in the Gemini app (type / and its name) or paste it as the first message.
• DeepSeek: through the DeepSeek API, use reasoning_effort low, or turn thinking off; this is a short writing job.
• Qwen: through Model Studio, thinking can be off (enable_thinking false) for this job.
• GLM: GLM-4.7-Flash is free on the Z.ai API. With GLM-5.3, use reasoning_effort low; requests that turn thinking off fail on that model.
• Kimi: in the Kimi app (kimi.com) you can paste several past WEEK FILEs at once and ask it to point out patterns.

Example

Input (illustrative). Last week's file: "Numbers: class seats sold 24 (target 30) · revenue €1,980 · online orders 3. Focus metric: class seats sold for the coming week, target 30, counted in the booking sheet on Friday. Protected time: photograph the new mug range for the online shop, Monday 14:00-16:00." Notes on this week: "A café down the road ordered 40 mugs after seeing them at the market. Seats sold 27, revenue €2,310. Online shop: 2 orders, that's the number that bugs me. About 5 hours redoing a glaze batch that came out patchy because I rushed the drying. 3-4 hours answering the same five class questions in DMs. Photos didn't happen, Monday went on class prep again. Mornings are quiet. Next week I just want the online shop to start working."

Output

Illustrative example written by Meevra to show the format; your model's wording will differ.

Troubleshooting

• It asks questions although you pasted notes. Make sure the notes come after the prompt in the same message, or reply "Use my notes and write the file now."
• It asks all five questions at once. Reply "One question per message; wait for my answer."
• A number you gave is missing. Reply "Put every figure I gave in Numbers, against last week and the target."
• It changed how last week's metric is counted. Reply "Keep last week's definition and unit for that metric."
• It reads like advice instead of your log. Reply "Write it as my log. Mark only next week's proposals with 'Proposed:'."
• The file contains things you did not say. Reply "Only use what I told you; remove anything else."
• It runs over 200 words. Reply "Cut to under 200 words; keep every heading and the Numbers line."

© 2026 Meevra · meevra.com · Licensed to the purchaser for personal and internal business use.