Your inbox sorted into replies and decisions
Unread count is 150. The actual decision count is five.
Tested Claude Opus 5.5, medium effort · 15 Nov 2026 · on the example values
Summarise messages first
Use subject, sender role, requested action and known deadline. No raw personal inbox.
Privacy Remove personal names where not needed, phone numbers, email IDs, personal records and bank details. Use role, employee code or aggregate business facts.
Whenever I send the next details, use this format. Triage rule: [Triage rule] Messages: [Messages] Reply tone: [Reply tone] A. Table with category and reason per message. B. Draft short (up to 6 lines) replies for decisions, no approval implied. C. Delegate list and missing info; do not send or change anything. Use only supplied facts and numbers. Separate facts from suggestions and label suggestions clearly. No invented reasons, commitments or figures. A person checks before sending, paying or deciding. Line limits count body lines only, not subject, greeting or sign-off. Keep every required fact; if they do not fit, say so instead of dropping one.
Set it up once. Save the approved report format and decision rules, not private records. Pick your tool at the top to see how.
What is that?
A Project is a folder in ChatGPT that keeps your instructions and files for every chat inside it. It's on the free plan too. Checked 14 Nov 2026.
What is that?
A Project in Claude is a workspace with its own instructions and files that every chat inside it uses. Checked 14 Nov 2026.
What is that?
A Gem is a custom Gemini that loads your instructions before every chat. Free on every plan; you create it on a computer. Checked 14 Nov 2026.
What is that?
Signed in with a work account, Microsoft's Copilot Chat applies enterprise data protection: your prompts aren't used to train its models. Checked 14 Nov 2026.
After that, each time: send the new approved facts and current rules; review before acting.
First time? Try it in a normal chat: paste and send.
Which model? The free one in any tool is fine.
In your AI tool: paste, send.
- M1Review price. Owner decision needed.
- M2Delegate. Stores checks.
- M4Confirm time. No time invented.
Then send this: Check all categories and draft facts, no unauthorised approval.
Sends reply
Reply: "Draft only."
Approves price
Reply: "Review needed."
Invents time
Reply: "Ask to confirm."
Sir ne bola urgent. Ab bolo: ho gaya.
Did it work?
If your whole team does this, we'll set it up for all of them in one session.