Deal with those old emails.
An old inbox mixes forgotten promises with conversations that finished somewhere else. Let AI reconstruct what still matters, prepare the work worth doing and clear away what no longer needs your attention.
Try it now
Copy the prompt into your AI tool of choice, then answer its questions as they come.
Paste it into any AI chat to start. For most people, this is all you need.
Everything on this page in one paste: the prompt plus the full plan, every step and what to check. Use it for longer moves, to plan ahead, to share with a colleague, or if your AI can't open links.
How it goes
What the AI will do
- Read bounded batches of old threads and search later replies, calendar events and task records to distinguish live obligations from resolved history.
- Draft useful replies, capture agreed tasks and archive categories you authorize.
- Verify completed changes and keep uncertain commitments or unsent drafts visible.
What to have ready
- The mailbox or folder and period you want to tackle; a rough sense of the backlog is enough.
- Access to the relevant email and, where available, calendar and task records, or an allowed export.
Where you come in
- Settle ambiguous commitments and confirm how categories of messages should be handled.
- Approve sending, deletion or new commitments when you want those actions taken.
What you'll get
A processed batch of old email, with agreed archives and task records checked, useful replies drafted, and a short list of decisions still needing you.
You're done when
The agreed batch has a checked disposition: resolved mail is archived where authorized, real work is captured without duplicates, useful replies are drafted or explicitly authorized and sent, and unresolved decisions are easy to find.
These are general instructions, not a description of this user's particular AI system, tools, access, or circumstances. Adapt the method to the user's goal and the capabilities you can actually use, while preserving the activity's essential sequence, review points, and intended result. ### Start with the activity and the user's context Read the complete activity and guide the user through it. Use the activity prompt yourself; the user does not need to paste it again. Keep the responsibilities of the user, the AI, and any colleague clear. Begin with what the user has already supplied. Establish the intended result and resolve missing context through focused questions when needed. Work within the activity's stated scope, permissions, required confirmations, and stopping points, including any deliberate order for introducing information. Make progress with the available information and reasonable, clearly identified assumptions where the activity allows them. ### Check the capabilities needed for this work Inspect relevant available tools, connectors, and data. Use small, relevant, non-destructive checks to establish whether necessary access or a capability actually works. Naming a product does not establish access to it. Let this activity determine what needs checking; do not impose a universal preliminary questionnaire or a full system audit. When useful access is missing: - Explain what you cannot reach and how that material would help. - Investigate practical setup, connection, export, or other access alternatives before asking the user to gather material or troubleshoot. - If progress depends on access, request only the specific user action or permission needed to enable it or provide the material another way. Continue useful independent work while awaiting the response. - Offer to walk the user through an export or upload. Welcome files, links, copied text, rough notes, or an interview as supported by your system and the activity, and use the best available route. If the missing access is optional, continue with what is available. ### Ask focused questions and respect the answers Ask questions only when necessary to complete the work well or satisfy the activity's required review points. Build on answers already supplied. Ask one question at a time, or a small group of closely related questions. Wait on work that depends on the answer, and continue useful work that does not. Explain unfamiliar language with examples where helpful. Confirm answers or decisions you supply on the user's behalf before treating them as the user's final answers or decisions. This does not require separate approval of every sourced finding; follow the activity's specific review requirements. ### Keep evidence and independent contributions distinct Distinguish evidence, inference, and missing information. Check your work against the source material. Use independent agents when your system actually supports delegation and independent investigation or review would improve the work: - Give each agent a bounded brief and the context appropriate to its role. - Keep discovery independent until synthesis where the activity requires it. Preserve any requirement for the user or colleagues to contribute before seeing other explanations. - Follow the activity's stated fallback and any required confirmation when independent agents or fresh contexts are unavailable. Preserve reviewer isolation where required; describe shared-context or sequential passes accurately. ### Produce, check, and revise the result Do the substantive investigation and production. Carry the activity through to the most complete, usable result within its scope, rather than ending with a plan or handing the main work back to the user. When missing information does not prevent useful work, proceed with clearly identified assumptions where the activity permits them. Check the result against the activity's completion criteria. At each substantive delivery, state what is complete and any material unfinished work, assumptions, helpful missing information, and remaining uncertainty. Explain how these affect what the user can use, and help the user revise the result. Preserve the activity's stopping point. Do not turn a rehearsal, investigation, or test-design activity into execution that the user has not authorized. Help me deal with a backlog of old email. Do the reading, follow-up investigation, drafting and agreed cleanup yourself. Keep me involved in decisions that require my judgment, not in sorting every message manually. ### Establish the batch and available access Identify the account, folder or search, date range and practical goal from what I supplied. If unclear, ask one focused question. Inspect the actual email connection or app and what it permits: search, complete thread reading, drafts, labels or folders, archive and task capture. Also inspect relevant calendar and task access. Try access before asking me to export or move material. Agree a manageable first batch based on the volume and complexity. Establish how archived mail remains retrievable and where drafts and tasks should go. Use existing instructions about handling messages, retention or categories; do not infer permission to delete or send from a request to clean up old mail. Record the batch search and its boundaries so we can continue without reopening finished work. If access is unavailable, investigate permitted setup or exports and help me provide a usable batch. An export may allow analysis and reply drafts without mailbox changes. Make that distinction early and continue with useful work; do not claim an exported message has been archived in the live inbox. ### Reconstruct what still needs action Read each selected conversation with enough surrounding messages to understand it. Separate requests to me, promises I made, items awaiting someone else, useful reference material and conversations that appear finished. Do not decide from subject lines or unread status alone. For anything that looks unfinished, search later replies, forwarded versions and related threads using the people, project, requested result and distinctive terms. Check relevant calendar events, meeting transcripts and task records for evidence that it was resolved elsewhere, superseded, cancelled or moved to another owner. Keep message dates, promised dates and actual event dates distinct. A meeting on the calendar does not establish that the promised deliverable was completed. Identify the recipient and relationship from actual correspondence when drafting would depend on tone, authority or history. Note their expressed concern and what our last exchange committed us to. Keep an inferred intention or likely reaction distinguishable from their words. For each live item, record a brief source-linked account: what was requested or promised, by whom, the best current evidence, any deadline that still applies and the smallest useful next action. Missing replies do not prove noncompletion. Ask about material uncertainty when records cannot resolve it; avoid presenting old proposals as accepted obligations. ### Prepare the useful work Draft overdue replies that respond to the current situation. Preserve my actual position and avoid inventing an excuse, a completed action or a new date. Use the source record to complete an answer or produce an attachment when the work is feasible within the agreed scope. If a reply depends on my choice, prepare what is known and isolate the single decision instead of filling it in for me. Create a proposed task for genuine work that remains, with a concrete result, source link and evidenced deadline. Do not silently assign it to a colleague or manufacture urgency. Check the existing task system first to avoid duplicating a task already recorded. A waiting-for item should identify whom we are waiting for and the evidence of that request. Suggest categories suitable for batch handling, such as resolved discussions or reference-only messages. Show representative examples and the rule so I can authorize the category without reviewing every email. Keep ambiguous messages out of a broad rule. Ask one focused question about the most useful unresolved decision while continuing unaffected preparation. ### Carry out the agreed cleanup After the relevant choices are settled, save drafts, create or update agreed tasks, and archive the authorized messages using actual supported actions. Respect authorization already given; do not repeatedly ask about the same category. Obtain explicit authorization before sending messages, deleting mail or making new commitments on my behalf. For a send, show the actual recipient, content and any attachment for approval unless that exact action is already authorized. Before mutation, retain the message identifiers and original location or labels needed to recover from an error. For moves or archives, inspect the resulting thread state. For tasks, verify that the correct destination contains one task with the right source and deadline. For authorized sends, check the sent record; for deletions, verify the actual result. If a tool response is uncertain, check current state before retrying to avoid duplicate tasks or messages. For a large backlog, divide bounded thread groups among agents for read-only investigation and draft suggestions where useful. Give them complete conversations, later-search access, classification rules and exact deliverables; keep all mailbox mutations under one coordinator. Track coverage and reconcile duplicate threads or commitments. Use successive batches or fresh-session handoffs when capacity is limited. ### Finish the batch and make continuation easy Deliver a short account of what changed: archived messages, saved drafts, captured live tasks and any authorized sends or deletions, using counts only from verified actions. Put the few decisions still needing me first, with direct links. Separate prepared work from completed mailbox actions. Check that every item in the batch has a disposition and that unresolved messages remain easy to find. Preserve the batch boundary and remaining search for continuation. Continue through the agreed scope, incorporating my answers as they arrive; do not let an endless incoming inbox silently enlarge the task. These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.
View full AI instructions
Deal with those old emails.
What the AI will do
- Read bounded batches of old threads and search later replies, calendar events and task records to distinguish live obligations from resolved history.
- Draft useful replies, capture agreed tasks and archive categories you authorize.
- Verify completed changes and keep uncertain commitments or unsent drafts visible.
What you can provide
- The mailbox or folder and period you want to tackle; a rough sense of the backlog is enough.
- Access to the relevant email and, where available, calendar and task records, or an allowed export.
Where you'll take part
- Settle ambiguous commitments and confirm how categories of messages should be handled.
- Approve sending, deletion or new commitments when you want those actions taken.
What you'll get
A processed batch of old email, with agreed archives and task records checked, useful replies drafted, and a short list of decisions still needing you.
Activity context
An old inbox mixes forgotten promises with conversations that finished somewhere else. Let AI reconstruct what still matters, prepare the work worth doing and clear away what no longer needs your attention.
For the user
Point the AI at the backlog you want to clear. It will work in manageable batches and look for later developments before treating an old email as unfinished. You decide uncertain commitments and authorize any sends or deletions; it carries out the agreed cleanup and reports what actually changed.
Guidance for the AI
These are general instructions, not a description of this user's particular AI system, tools, access, or circumstances. Adapt the method to the user's goal and the capabilities you can actually use, while preserving the activity's essential sequence, review points, and intended result.
Start with the activity and the user's context
Read the complete activity and guide the user through it. Use the activity prompt yourself; the user does not need to paste it again. Keep the responsibilities of the user, the AI, and any colleague clear.
Begin with what the user has already supplied. Establish the intended result and resolve missing context through focused questions when needed. Work within the activity's stated scope, permissions, required confirmations, and stopping points, including any deliberate order for introducing information. Make progress with the available information and reasonable, clearly identified assumptions where the activity allows them.
Check the capabilities needed for this work
Inspect relevant available tools, connectors, and data. Use small, relevant, non-destructive checks to establish whether necessary access or a capability actually works. Naming a product does not establish access to it. Let this activity determine what needs checking; do not impose a universal preliminary questionnaire or a full system audit.
When useful access is missing:
- Explain what you cannot reach and how that material would help.
- Investigate practical setup, connection, export, or other access alternatives before asking the user to gather material or troubleshoot.
- If progress depends on access, request only the specific user action or permission needed to enable it or provide the material another way. Continue useful independent work while awaiting the response.
- Offer to walk the user through an export or upload. Welcome files, links, copied text, rough notes, or an interview as supported by your system and the activity, and use the best available route. If the missing access is optional, continue with what is available.
Ask focused questions and respect the answers
Ask questions only when necessary to complete the work well or satisfy the activity's required review points. Build on answers already supplied. Ask one question at a time, or a small group of closely related questions. Wait on work that depends on the answer, and continue useful work that does not. Explain unfamiliar language with examples where helpful.
Confirm answers or decisions you supply on the user's behalf before treating them as the user's final answers or decisions. This does not require separate approval of every sourced finding; follow the activity's specific review requirements.
Keep evidence and independent contributions distinct
Distinguish evidence, inference, and missing information. Check your work against the source material.
Use independent agents when your system actually supports delegation and independent investigation or review would improve the work:
- Give each agent a bounded brief and the context appropriate to its role.
- Keep discovery independent until synthesis where the activity requires it. Preserve any requirement for the user or colleagues to contribute before seeing other explanations.
- Follow the activity's stated fallback and any required confirmation when independent agents or fresh contexts are unavailable. Preserve reviewer isolation where required; describe shared-context or sequential passes accurately.
Produce, check, and revise the result
Do the substantive investigation and production. Carry the activity through to the most complete, usable result within its scope, rather than ending with a plan or handing the main work back to the user. When missing information does not prevent useful work, proceed with clearly identified assumptions where the activity permits them.
Check the result against the activity's completion criteria. At each substantive delivery, state what is complete and any material unfinished work, assumptions, helpful missing information, and remaining uncertainty. Explain how these affect what the user can use, and help the user revise the result.
Preserve the activity's stopping point. Do not turn a rehearsal, investigation, or test-design activity into execution that the user has not authorized.
Your AI prompt
Help me deal with a backlog of old email. Do the reading, follow-up investigation, drafting and agreed cleanup yourself. Keep me involved in decisions that require my judgment, not in sorting every message manually.
Establish the batch and available access
Identify the account, folder or search, date range and practical goal from what I supplied. If unclear, ask one focused question. Inspect the actual email connection or app and what it permits: search, complete thread reading, drafts, labels or folders, archive and task capture. Also inspect relevant calendar and task access. Try access before asking me to export or move material.
Agree a manageable first batch based on the volume and complexity. Establish how archived mail remains retrievable and where drafts and tasks should go. Use existing instructions about handling messages, retention or categories; do not infer permission to delete or send from a request to clean up old mail. Record the batch search and its boundaries so we can continue without reopening finished work.
If access is unavailable, investigate permitted setup or exports and help me provide a usable batch. An export may allow analysis and reply drafts without mailbox changes. Make that distinction early and continue with useful work; do not claim an exported message has been archived in the live inbox.
Reconstruct what still needs action
Read each selected conversation with enough surrounding messages to understand it. Separate requests to me, promises I made, items awaiting someone else, useful reference material and conversations that appear finished. Do not decide from subject lines or unread status alone.
For anything that looks unfinished, search later replies, forwarded versions and related threads using the people, project, requested result and distinctive terms. Check relevant calendar events, meeting transcripts and task records for evidence that it was resolved elsewhere, superseded, cancelled or moved to another owner. Keep message dates, promised dates and actual event dates distinct. A meeting on the calendar does not establish that the promised deliverable was completed.
Identify the recipient and relationship from actual correspondence when drafting would depend on tone, authority or history. Note their expressed concern and what our last exchange committed us to. Keep an inferred intention or likely reaction distinguishable from their words.
For each live item, record a brief source-linked account: what was requested or promised, by whom, the best current evidence, any deadline that still applies and the smallest useful next action. Missing replies do not prove noncompletion. Ask about material uncertainty when records cannot resolve it; avoid presenting old proposals as accepted obligations.
Prepare the useful work
Draft overdue replies that respond to the current situation. Preserve my actual position and avoid inventing an excuse, a completed action or a new date. Use the source record to complete an answer or produce an attachment when the work is feasible within the agreed scope. If a reply depends on my choice, prepare what is known and isolate the single decision instead of filling it in for me.
Create a proposed task for genuine work that remains, with a concrete result, source link and evidenced deadline. Do not silently assign it to a colleague or manufacture urgency. Check the existing task system first to avoid duplicating a task already recorded. A waiting-for item should identify whom we are waiting for and the evidence of that request.
Suggest categories suitable for batch handling, such as resolved discussions or reference-only messages. Show representative examples and the rule so I can authorize the category without reviewing every email. Keep ambiguous messages out of a broad rule. Ask one focused question about the most useful unresolved decision while continuing unaffected preparation.
Carry out the agreed cleanup
After the relevant choices are settled, save drafts, create or update agreed tasks, and archive the authorized messages using actual supported actions. Respect authorization already given; do not repeatedly ask about the same category. Obtain explicit authorization before sending messages, deleting mail or making new commitments on my behalf. For a send, show the actual recipient, content and any attachment for approval unless that exact action is already authorized.
Before mutation, retain the message identifiers and original location or labels needed to recover from an error. For moves or archives, inspect the resulting thread state. For tasks, verify that the correct destination contains one task with the right source and deadline. For authorized sends, check the sent record; for deletions, verify the actual result. If a tool response is uncertain, check current state before retrying to avoid duplicate tasks or messages.
For a large backlog, divide bounded thread groups among agents for read-only investigation and draft suggestions where useful. Give them complete conversations, later-search access, classification rules and exact deliverables; keep all mailbox mutations under one coordinator. Track coverage and reconcile duplicate threads or commitments. Use successive batches or fresh-session handoffs when capacity is limited.
Finish the batch and make continuation easy
Deliver a short account of what changed: archived messages, saved drafts, captured live tasks and any authorized sends or deletions, using counts only from verified actions. Put the few decisions still needing me first, with direct links. Separate prepared work from completed mailbox actions.
Check that every item in the batch has a disposition and that unresolved messages remain easy to find. Preserve the batch boundary and remaining search for continuation. Continue through the agreed scope, incorporating my answers as they arrive; do not let an endless incoming inbox silently enlarge the task.
Done when
The agreed batch has a checked disposition: resolved mail is archived where authorized, real work is captured without duplicates, useful replies are drafted or explicitly authorized and sent, and unresolved decisions are easy to find.
These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.
These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.