Overview
A weekly review is the highest-leverage 30 minutes in a knowledge worker's week. This prompt-chain gives ChatGPT the raw material — completed events, tasks, and free-form notes — and returns a structured review you can actually act on.
How it works
- You paste three inputs: last week's completed calendar events, closed tasks, and any notes.
- Stage 1 prompt extracts wins, blockers, and unresolved threads.
- Stage 2 prompt asks 'what would a great manager coach you on?' and returns 3 questions.
- Stage 3 prompt drafts a 5-item priority list for next week with time estimates.
Benefits
- Turns fragmented data into a single narrative you can share with a manager.
- Coaching questions force self-reflection without hiring a coach.
- Time-estimated priorities calibrate what's realistic before Monday.
Use cases
- IC engineers preparing weekly updates.
- Founders doing solo Friday retros.
- Managers preparing 1:1s with themselves before their skip-level.
Step-by-step guide
Step 1: Export your week
Copy completed events from Calendar, closed tasks from your task manager, and any notes from Notes/Notion.
Step 2: Run stage 1
Paste inputs with the extraction prompt. Ask for a table: category, item, outcome, effort.
Step 3: Run stage 2
Feed stage 1 output back and ask for coaching questions. Answer them in your own words before continuing.
Step 4: Run stage 3
Ask for next-week priorities based on stage 1 + your answers. Cap at 5 with time estimates that sum to ~60% of your available focus time.
Step 5: Save the output
Store the review in a running doc. Over months it becomes a career-review dataset.