Build a bot that finds meetings to eliminate.
A calendar fills up one recurring invitation at a time. Build a reusable reviewer that checks whether those meetings still earn their place—and describes the AI work that could replace the parts that don't. The useful result is a repeatable way to notice opportunities, with enough detail to act on a good one.
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
- Examine several occurrences of recurring meetings, including their agendas, conversations, decisions and follow-up work, to find work AI could take over.
- Create and try a reusable meeting-review prompt or supported workflow that reports new candidates and changes since its last review.
- When there is a promising candidate you want to pursue, describe a concrete replacement tool. If you request it, prepare a complete prompt for building that tool and help test the result.
What to have ready
- A team or calendar to examine and access to useful invitations, agendas, transcripts and follow-up messages. A few pasted or exported meeting records can support a smaller review.
- Any meeting review or replacement work you already have, so the AI can build on it.
Where you come in
- Correct what the AI thinks the meetings accomplish and choose the scope and way you want to rerun the review.
- Review a sample report and decide which replacement, if any, to pursue.
What you'll get
You receive a reusable reviewer and a sample report, plus a concrete AI replacement proposal when a suitable candidate is selected. The report can find that no current meeting merits replacement. Delivery explains whether the reviewer runs on demand, is actually scheduled, or still needs setup.
You're done when
You have a reusable meeting reviewer, a sample report and clear instructions for running it again, with tested and untested capabilities distinguished. If you selected a suitable meeting, you also have its replacement proposal and any requested build prompt; a supported finding of no current candidates is a complete result.
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. ### 1. Establish what the reviewer will examine Help me build a repeatable way to find recurring meetings that could be eliminated or shortened by having AI perform specific work. If I already have a meeting audit, saved prompt or replacement prototype, inspect it and preserve useful work. Ask which team or calendar to cover only if the context does not establish it. Inspect and try relevant calendar, email, company chat and other message access, including WhatsApp where available, document access and meeting transcripts. Use meeting names, organizers, attendees and linked projects to find invitations, agendas, recordings or transcripts, preparation, decisions, action items and follow-up exchanges across several occurrences. Follow linked project records to see whether meeting outputs were used. State the period and actual coverage. If access is missing, inspect available connection or export options before asking me for a specific action. Explain which records would unlock the review and offer to draft a focused IT access request if useful. Work with pasted or exported invitations and several meeting records when available. If only recollection is possible, interview me one question at a time and label the result a provisional reviewer tested on my account; it cannot establish actual meeting histories or autonomously scan a calendar. ### 2. Distinguish the meeting's jobs For each meeting series, trace what happens before, during and after it. Identify information collection, status sharing, exception resolution, decisions, coordination, relationship work and other purposes supported by the record. Establish attendees' actual roles and expressed needs from invitations, correspondence and transcripts; attendance alone does not tell you why someone needs the meeting. Look for work that recurs predictably, duplicates available information or produces little subsequent action. Also look for value a transcript can miss: a decision that needed discussion, a conflict resolved, or a dependency kept visible. Keep a lack of evidence distinct from evidence that the meeting is unnecessary. Compare multiple occurrences before proposing a pattern. For a large calendar, divide complete meeting series among agents, with explicit source lists, date ranges and a common reporting format. Have each return the meeting's jobs, evidence, replacement possibilities, dependencies and unknowns. Reconcile duplicated invitations and shared work centrally. If agents are unavailable, use bounded passes and maintain the same coverage record. Present a few promising candidates with sources, the work AI could do, the human conversation that would remain, and a supported estimate of possible time released. Distinguish invite duration from actual attendance and any estimated preparation effort. Ask me to correct the account and choose a candidate for a detailed replacement proposal. If the evidence yields no suitable candidate, or I choose none, retain that result without manufacturing a recommendation. Continue to build and test the reusable reviewer; its sample report can say “no current candidates” with the evidence and coverage behind that conclusion. Skip the candidate-specific replacement and tool-build phases until there is a supported candidate I want to pursue. ### 3. Specify a concrete replacement If I selected a supported candidate, describe an actual tool or workflow. A weekly exception briefing might read project updates, reconcile conflicting status claims, flag missing evidence and prepare the few questions that need a decision. Another meeting might need a tool that collects and reconciles updates, or maintains an answered-and-open question record. Choose from the meeting's actual work. Specify its inputs and where they come from, when they become available, transformations and checks, output format and destination, decisions reserved for people, and what it does when a source is absent or contradictory. Show a small example output using accessible evidence, clearly marking hypothetical data if needed. Explain what remaining agenda this would leave and which participant needs it would serve. This is the replacement proposal; creating the replacement tool is an optional later phase. ### 4. Build the reusable reviewer Inspect the actual AI environment and current official documentation for ways it can save and rerun this review. Determine whether it supports a saved reusable prompt or skill, a scheduled task with access to the required records, or an approved automation. Check execution, data access and output delivery separately: a reminder to do the review is not a review that runs automatically. Recommend the simplest supported arrangement that fits my intended cadence. Explain what starts each run, what records it reads, where results appear and any setup I must do. Ask for my choice if the arrangement is unresolved; use a choice I have already made. Obtain my decision before enabling a schedule or external integration. Create the reusable review instructions and configure the agreed arrangement where authorized and possible. The reviewer must specify the scope and review period; retrieve the relevant meeting and follow-up records; compare with the previous report; identify new or changed candidates; show evidence, missing records and proposed replacement work; and retain prior decisions so it does not recommend the same rejected change without new evidence. Include where that prior report or decision record will be stored and how a future run retrieves it. If persistence is unavailable, make attaching the previous report an explicit input. Run it on a small real sample. Check source coverage, whether its recommendations preserve actual meeting work, and whether its report makes the next human decision clear. Use a separate reviewer with the report and original meeting evidence where available; otherwise check in a separate pass and disclose shared context. Fix substantive errors. Rerun against the same sample with a prior report attached to check that it recognizes unchanged findings; test a later period if available without pretending that historical replay proves unattended operation. If configuration or execution is unavailable, deliver the complete portable prompt with its required inputs and precise run instructions. Label it as the on-demand fallback and state what you could and could not test. Do not describe a suggested schedule as installed. ### 5. Offer the selected tool's build handoff and finish If I request a prompt to create the meeting replacement tool, write a self-contained build prompt incorporating the evidence already gathered, purpose, users, agreed inputs and connections, processing steps, example output, human decision boundaries, missing-data behavior, representative test inputs and acceptance criteria. Include relevant actual environment constraints and the intended delivery point. Do not make me reconstruct this conversation. Help me use that prompt in an environment with the necessary capabilities. If I ask you to build the tool and you can, perform the build, test it on representative inputs, compare the result with the meeting's actual job and repair failures. Otherwise supply the ready-to-use handoff and identify the specific remaining setup. Deliver the reusable reviewer and sample report, plus a replacement specification when a candidate was selected and any requested build prompt or tested tool. A supported no-candidate finding is a complete result. Explain how to rerun it and what is configured versus proposed. Leave calendar changes, attendee messages and cancellation decisions to me unless separately authorized. 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
Build a bot that finds meetings to eliminate.
What the AI will do
- Examine several occurrences of recurring meetings, including their agendas, conversations, decisions and follow-up work, to find work AI could take over.
- Create and try a reusable meeting-review prompt or supported workflow that reports new candidates and changes since its last review.
- When there is a promising candidate you want to pursue, describe a concrete replacement tool. If you request it, prepare a complete prompt for building that tool and help test the result.
What you can provide
- A team or calendar to examine and access to useful invitations, agendas, transcripts and follow-up messages. A few pasted or exported meeting records can support a smaller review.
- Any meeting review or replacement work you already have, so the AI can build on it.
Where you'll take part
- Correct what the AI thinks the meetings accomplish and choose the scope and way you want to rerun the review.
- Review a sample report and decide which replacement, if any, to pursue.
What you'll get
You receive a reusable reviewer and a sample report, plus a concrete AI replacement proposal when a suitable candidate is selected. The report can find that no current meeting merits replacement. Delivery explains whether the reviewer runs on demand, is actually scheduled, or still needs setup.
Activity context
A calendar fills up one recurring invitation at a time. Build a reusable reviewer that checks whether those meetings still earn their place—and describes the AI work that could replace the parts that don't. The useful result is a repeatable way to notice opportunities, with enough detail to act on a good one.
For the user
Start with the team or recurring meetings you want to examine; rough context is enough. The AI will investigate available records and the tools in your actual setup. You will check its understanding, choose how to rerun the review, and review its first result. You can then ask for a complete prompt to build a proposed meeting replacement.
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
1. Establish what the reviewer will examine
Help me build a repeatable way to find recurring meetings that could be eliminated or shortened by having AI perform specific work. If I already have a meeting audit, saved prompt or replacement prototype, inspect it and preserve useful work. Ask which team or calendar to cover only if the context does not establish it.
Inspect and try relevant calendar, email, company chat and other message access, including WhatsApp where available, document access and meeting transcripts. Use meeting names, organizers, attendees and linked projects to find invitations, agendas, recordings or transcripts, preparation, decisions, action items and follow-up exchanges across several occurrences. Follow linked project records to see whether meeting outputs were used. State the period and actual coverage.
If access is missing, inspect available connection or export options before asking me for a specific action. Explain which records would unlock the review and offer to draft a focused IT access request if useful. Work with pasted or exported invitations and several meeting records when available. If only recollection is possible, interview me one question at a time and label the result a provisional reviewer tested on my account; it cannot establish actual meeting histories or autonomously scan a calendar.
2. Distinguish the meeting's jobs
For each meeting series, trace what happens before, during and after it. Identify information collection, status sharing, exception resolution, decisions, coordination, relationship work and other purposes supported by the record. Establish attendees' actual roles and expressed needs from invitations, correspondence and transcripts; attendance alone does not tell you why someone needs the meeting.
Look for work that recurs predictably, duplicates available information or produces little subsequent action. Also look for value a transcript can miss: a decision that needed discussion, a conflict resolved, or a dependency kept visible. Keep a lack of evidence distinct from evidence that the meeting is unnecessary. Compare multiple occurrences before proposing a pattern.
For a large calendar, divide complete meeting series among agents, with explicit source lists, date ranges and a common reporting format. Have each return the meeting's jobs, evidence, replacement possibilities, dependencies and unknowns. Reconcile duplicated invitations and shared work centrally. If agents are unavailable, use bounded passes and maintain the same coverage record.
Present a few promising candidates with sources, the work AI could do, the human conversation that would remain, and a supported estimate of possible time released. Distinguish invite duration from actual attendance and any estimated preparation effort. Ask me to correct the account and choose a candidate for a detailed replacement proposal. If the evidence yields no suitable candidate, or I choose none, retain that result without manufacturing a recommendation. Continue to build and test the reusable reviewer; its sample report can say “no current candidates” with the evidence and coverage behind that conclusion. Skip the candidate-specific replacement and tool-build phases until there is a supported candidate I want to pursue.
3. Specify a concrete replacement
If I selected a supported candidate, describe an actual tool or workflow. A weekly exception briefing might read project updates, reconcile conflicting status claims, flag missing evidence and prepare the few questions that need a decision. Another meeting might need a tool that collects and reconciles updates, or maintains an answered-and-open question record. Choose from the meeting's actual work.
Specify its inputs and where they come from, when they become available, transformations and checks, output format and destination, decisions reserved for people, and what it does when a source is absent or contradictory. Show a small example output using accessible evidence, clearly marking hypothetical data if needed. Explain what remaining agenda this would leave and which participant needs it would serve. This is the replacement proposal; creating the replacement tool is an optional later phase.
4. Build the reusable reviewer
Inspect the actual AI environment and current official documentation for ways it can save and rerun this review. Determine whether it supports a saved reusable prompt or skill, a scheduled task with access to the required records, or an approved automation. Check execution, data access and output delivery separately: a reminder to do the review is not a review that runs automatically.
Recommend the simplest supported arrangement that fits my intended cadence. Explain what starts each run, what records it reads, where results appear and any setup I must do. Ask for my choice if the arrangement is unresolved; use a choice I have already made. Obtain my decision before enabling a schedule or external integration.
Create the reusable review instructions and configure the agreed arrangement where authorized and possible. The reviewer must specify the scope and review period; retrieve the relevant meeting and follow-up records; compare with the previous report; identify new or changed candidates; show evidence, missing records and proposed replacement work; and retain prior decisions so it does not recommend the same rejected change without new evidence. Include where that prior report or decision record will be stored and how a future run retrieves it. If persistence is unavailable, make attaching the previous report an explicit input.
Run it on a small real sample. Check source coverage, whether its recommendations preserve actual meeting work, and whether its report makes the next human decision clear. Use a separate reviewer with the report and original meeting evidence where available; otherwise check in a separate pass and disclose shared context. Fix substantive errors. Rerun against the same sample with a prior report attached to check that it recognizes unchanged findings; test a later period if available without pretending that historical replay proves unattended operation.
If configuration or execution is unavailable, deliver the complete portable prompt with its required inputs and precise run instructions. Label it as the on-demand fallback and state what you could and could not test. Do not describe a suggested schedule as installed.
5. Offer the selected tool's build handoff and finish
If I request a prompt to create the meeting replacement tool, write a self-contained build prompt incorporating the evidence already gathered, purpose, users, agreed inputs and connections, processing steps, example output, human decision boundaries, missing-data behavior, representative test inputs and acceptance criteria. Include relevant actual environment constraints and the intended delivery point. Do not make me reconstruct this conversation.
Help me use that prompt in an environment with the necessary capabilities. If I ask you to build the tool and you can, perform the build, test it on representative inputs, compare the result with the meeting's actual job and repair failures. Otherwise supply the ready-to-use handoff and identify the specific remaining setup.
Deliver the reusable reviewer and sample report, plus a replacement specification when a candidate was selected and any requested build prompt or tested tool. A supported no-candidate finding is a complete result. Explain how to rerun it and what is configured versus proposed. Leave calendar changes, attendee messages and cancellation decisions to me unless separately authorized.
Done when
You have a reusable meeting reviewer, a sample report and clear instructions for running it again, with tested and untested capabilities distinguished. If you selected a suitable meeting, you also have its replacement proposal and any requested build prompt; a supported finding of no current candidates is a complete result.
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.