Ask one model to critique another’s answer.
A second model may notice what the first missed. Have it make a substantive critique, then check which criticisms deserve to change the answer.
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
- Have another model assess the original answer against the task and evidence, then evaluate the criticism and produce a checked revision.
What to have ready
- The original task and answer; available sources and connected project context can complete the review brief.
Where you come in
- Clarify what a good answer requires; if direct access is unavailable, paste the prepared brief into another model and bring back its response.
What you'll get
You have a revised answer informed by an actual second model’s critique, with the changes checked and unresolved disagreements explained—or a clearly identified partial result or prepared handoff if model identity or access cannot be verified.
You're done when
You have a revised answer informed by an actual second model’s critique, with the changes checked and unresolved disagreements explained—or a clearly identified partial result or prepared handoff if model identity or access cannot be verified.
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. Have a different model critique this answer, assess which criticisms are justified, and produce a better final answer. ### 1. Reconstruct the task and the available evidence. Read the original request, answer, and any sources or prior constraints I supplied. Establish what a successful answer should accomplish and distinguish factual correctness, reasoning, completeness, usefulness, and presentation where relevant. Use relevant available emails, messages and chats, meeting transcripts, documents, and project files to understand the audience, actual decision, and constraints. Gather enough source material for another model to evaluate the answer rather than judging its confidence or style alone. Ask focused questions when my intended result or missing context would change the assessment. If two review emphases would be useful, such as a factual check or a challenge to the recommendation’s reasoning, explain their difference and help me choose. A broad review may include both when appropriate; avoid a fixed questionnaire where the task is already clear. ### 2. Give a different model a complete review brief. Check which models you can actually invoke, and determine which produced the first answer where known. Use a genuinely different model for the second opinion. A new chat or another agent using the same model is not evidence of a different-model comparison. If the first model is unknown, ask me or inspect available provenance before claiming the reviewer is a different model. If its identity cannot be established, a critique can still be useful, but label the model distinction unverified and the requested different-model comparison partial. Do not claim that requirement has been met. Prepare a self-contained brief containing the original task, intended result, constraints, the exact first answer, and the relevant evidence. Include readable excerpts or attachments where a reviewer cannot access a source link. Clearly label any shortened or omitted context. Ask the reviewer to identify specific errors, unsupported assumptions, missing work, and useful improvements, citing the answer and evidence; invite it to conclude that no substantive change is needed. Invoke the second model if a suitable tool is available. Otherwise investigate a practical supported route, then provide the complete brief for me to paste into a model I can use. Ask me to return its actual response and wait before proceeding as though a critique exists. Do not fabricate that response or substitute a same-model self-review without explicitly agreeing a different, partial activity. ### 3. Evaluate the critique and revise. Read the returned critique against the task and original sources. Verify factual objections and test consequential reasoning where available tools permit. A critic can be wrong; distinguish supported corrections, useful alternative judgments, and unsupported objections. Follow up on material uncertainty rather than accepting every criticism automatically. If the critique exposes a choice of goals or values only I can make, explain two concrete alternatives where helpful, ask, and wait before adopting a position on my behalf. Otherwise revise the answer directly using justified changes, retaining strengths of the original. Where the original is substantively sound, a checked unchanged answer is acceptable; do not manufacture disagreement. ### 4. Deliver the checked result. - Check the revised answer against the agreed criteria and source evidence. - Present the usable answer first, then briefly explain important changes or why a major criticism was not adopted. - Name the models and describe the actual review setup where known, including missing context or access that affects the comparison. If another model’s response remains unavailable, leave the ready-to-paste brief and exact next step, clearly stating that the requested second-model review is unfinished. If you obtained a critique but cannot verify the two model identities differ, deliver the useful revision while retaining that explicit partial-comparison status. Do not label the original answer as independently checked. The intended result is an answer improved or confirmed by an actual second model, not simply two simulated viewpoints. 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
Ask one model to critique another’s answer.
What the AI will do
- Have another model assess the original answer against the task and evidence, then evaluate the criticism and produce a checked revision.
What you can provide
- The original task and answer; available sources and connected project context can complete the review brief.
Where you'll take part
- Clarify what a good answer requires; if direct access is unavailable, paste the prepared brief into another model and bring back its response.
What you'll get
You have a revised answer informed by an actual second model’s critique, with the changes checked and unresolved disagreements explained—or a clearly identified partial result or prepared handoff if model identity or access cannot be verified.
Activity context
A second model may notice what the first missed. Have it make a substantive critique, then check which criticisms deserve to change the answer.
For the user
Share the original task and answer. The AI will gather the context a reviewer needs and use a genuinely different model where available. If it cannot reach one, it can prepare a complete brief for you to paste into your other model and work with the returned critique.
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
Have a different model critique this answer, assess which criticisms are justified, and produce a better final answer.
1. Reconstruct the task and the available evidence.
Read the original request, answer, and any sources or prior constraints I supplied. Establish what a successful answer should accomplish and distinguish factual correctness, reasoning, completeness, usefulness, and presentation where relevant.
Use relevant available emails, messages and chats, meeting transcripts, documents, and project files to understand the audience, actual decision, and constraints. Gather enough source material for another model to evaluate the answer rather than judging its confidence or style alone. Ask focused questions when my intended result or missing context would change the assessment.
If two review emphases would be useful, such as a factual check or a challenge to the recommendation’s reasoning, explain their difference and help me choose. A broad review may include both when appropriate; avoid a fixed questionnaire where the task is already clear.
2. Give a different model a complete review brief.
Check which models you can actually invoke, and determine which produced the first answer where known. Use a genuinely different model for the second opinion. A new chat or another agent using the same model is not evidence of a different-model comparison. If the first model is unknown, ask me or inspect available provenance before claiming the reviewer is a different model. If its identity cannot be established, a critique can still be useful, but label the model distinction unverified and the requested different-model comparison partial. Do not claim that requirement has been met.
Prepare a self-contained brief containing the original task, intended result, constraints, the exact first answer, and the relevant evidence. Include readable excerpts or attachments where a reviewer cannot access a source link. Clearly label any shortened or omitted context. Ask the reviewer to identify specific errors, unsupported assumptions, missing work, and useful improvements, citing the answer and evidence; invite it to conclude that no substantive change is needed.
Invoke the second model if a suitable tool is available. Otherwise investigate a practical supported route, then provide the complete brief for me to paste into a model I can use. Ask me to return its actual response and wait before proceeding as though a critique exists. Do not fabricate that response or substitute a same-model self-review without explicitly agreeing a different, partial activity.
3. Evaluate the critique and revise.
Read the returned critique against the task and original sources. Verify factual objections and test consequential reasoning where available tools permit. A critic can be wrong; distinguish supported corrections, useful alternative judgments, and unsupported objections. Follow up on material uncertainty rather than accepting every criticism automatically.
If the critique exposes a choice of goals or values only I can make, explain two concrete alternatives where helpful, ask, and wait before adopting a position on my behalf. Otherwise revise the answer directly using justified changes, retaining strengths of the original. Where the original is substantively sound, a checked unchanged answer is acceptable; do not manufacture disagreement.
4. Deliver the checked result.
- Check the revised answer against the agreed criteria and source evidence.
- Present the usable answer first, then briefly explain important changes or why a major criticism was not adopted.
- Name the models and describe the actual review setup where known, including missing context or access that affects the comparison.
If another model’s response remains unavailable, leave the ready-to-paste brief and exact next step, clearly stating that the requested second-model review is unfinished. If you obtained a critique but cannot verify the two model identities differ, deliver the useful revision while retaining that explicit partial-comparison status. Do not label the original answer as independently checked. The intended result is an answer improved or confirmed by an actual second model, not simply two simulated viewpoints.
Done when
You have a revised answer informed by an actual second model’s critique, with the changes checked and unresolved disagreements explained—or a clearly identified partial result or prepared handoff if model identity or access cannot be verified.
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.