Find the original source behind a confusing claim.
Find what the claim rests on, and what changed along the way.
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
- Follow a claim through messages, documents, citations, and available underlying data.
- Check the source in context and explain what it supports and what changed in retelling.
What to have ready
- The claim and where you encountered it: a message, document, slide, link, screenshot, or rough account.
- Access to the relevant email, chat, document, meeting, and data trail is central; an interview alone can suggest leads but may not establish the origin.
Where you come in
- Clarify which claim matters and help resolve inaccessible references or ambiguous wording.
What you'll get
You have a traceable account of the claim and the evidence behind it, including changes in retelling and any unresolved origin or missing links.
You're done when
You have a traceable account of the claim and the evidence behind it, including changes in retelling and any unresolved origin or missing links.
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 find the original source behind a confusing claim. Trace what was said, where it came from, and whether the underlying evidence supports the version I encountered. ### 1. Identify the claim and establish access to its trail Read the wording and surrounding context I provide. Identify the particular factual assertion to trace; separate multiple claims if a sentence bundles them. Preserve exact wording, relevant numbers, dates, and qualifiers. Ask a focused question if the intended claim or its starting location remains ambiguous. Inspect available access and search relevant emails, chats and other messages, documents and their versions, calendar-linked meetings and transcripts, project records, company data, and public citations where relevant. Use distinctive phrases, names, quantities, links, and dates from the claim to locate related versions. Look at the surrounding exchange to understand whether someone was stating a finding, asking a question, repeating a rumor, or offering a hypothetical. Access to that data and message trail is central. If it is missing, investigate connection and export options and help obtain the specific original thread, document, recording, or dataset. Say upfront that reconstructing the trail through an interview will be difficult and incomplete. My recollection may suggest search leads, but it cannot by itself establish provenance or the original source. Continue useful accessible branches and accurately mark inaccessible ones. ### 2. Follow the links backward without losing the variants Build a source trail, keeping references sufficient to reopen each item. For each occurrence, record: - Its actual wording and qualifiers, author or publisher where known, and publication or message date. - The source it cites, links, attaches, or appears to paraphrase, distinguishing explicit attribution from your inference. - Changes to scope, population, timeframe, certainty, quantities, or meaning relative to nearby versions. Inspect linked originals and attachments. Follow citations through summaries to the document, data, interview, or other evidence behind them. Compare dates carefully: the earliest item found is not necessarily the origin, and a modified file date is not a publication date. Detect circular citations and repeated unsourced versions rather than counting them as independent confirmation. Use archived versions where accessible and needed, marking which version was examined. When the trail branches, assign separate bounded agents to different citation paths, message threads, or archives where available. Give each the precise claim, known starting source, search scope, and a requirement to return actual inspected evidence with unresolved links. Let first searches finish before reconciling them. Keep a coverage record and use waves or compact fresh-session handoffs when the trail exceeds one context. Without subagents, investigate separate branches in manageable passes and describe their independence accurately. ### 3. Check the underlying evidence in context Read enough of the strongest available underlying source to establish what it actually says. Examine definitions, methods, sample or population, dates, denominators, and limitations that affect this claim. If the claim rests on company data, inspect the relevant calculation, reporting period, and version where available. Recalculate a material number when the underlying data permits; report when it does not. Have a separate verifier inspect the candidate source and compare it with the claim. Ask whether the source directly supports the wording, supports only a narrower statement, contradicts it, or cannot resolve it. Give the verifier the sources and claim, not an instruction to validate the lead's preferred conclusion. Resolve disagreements by revisiting evidence. If a separate verifier is unavailable, do a distinct source-verification pass after discovery, reopening the candidate evidence and comparing it with the exact claim. Use a fresh-session handoff where possible; disclose when verification shares discovery context. A source can be the actual origin of a claim without proving that claim true. ### 4. Deliver the finding and a usable correction Lead with what the evidence supports, followed by the traceable path and material changes in retelling. Distinguish a verified original source, the earliest accessible occurrence, and an unresolved origin. Include direct links or usable source references and short quotations only where they help make a consequential difference clear. Offer a corrected, properly qualified formulation I can use, preserving unresolved details rather than filling them in. If the trail runs out, state exactly where and what missing evidence would be needed to continue. A transparent unresolved finding is a valid investigation result; do not promote repeated claims or an interview reconstruction into verified provenance. Stop with the source account and proposed wording, without contacting message authors or changing shared documents. 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
Find the original source behind a confusing claim.
What the AI will do
- Follow a claim through messages, documents, citations, and available underlying data.
- Check the source in context and explain what it supports and what changed in retelling.
What you can provide
- The claim and where you encountered it: a message, document, slide, link, screenshot, or rough account.
- Access to the relevant email, chat, document, meeting, and data trail is central; an interview alone can suggest leads but may not establish the origin.
Where you'll take part
- Clarify which claim matters and help resolve inaccessible references or ambiguous wording.
What you'll get
You have a traceable account of the claim and the evidence behind it, including changes in retelling and any unresolved origin or missing links.
Activity context
Find what the claim rests on, and what changed along the way.
For the user
Share the claim and its surrounding message or document if you have it. The AI will follow the actual trail and check the evidence. If that trail is inaccessible, it will explain why reconstructing it from memory is difficult and identify which conclusions remain unresolved.
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 find the original source behind a confusing claim. Trace what was said, where it came from, and whether the underlying evidence supports the version I encountered.
1. Identify the claim and establish access to its trail
Read the wording and surrounding context I provide. Identify the particular factual assertion to trace; separate multiple claims if a sentence bundles them. Preserve exact wording, relevant numbers, dates, and qualifiers. Ask a focused question if the intended claim or its starting location remains ambiguous.
Inspect available access and search relevant emails, chats and other messages, documents and their versions, calendar-linked meetings and transcripts, project records, company data, and public citations where relevant. Use distinctive phrases, names, quantities, links, and dates from the claim to locate related versions. Look at the surrounding exchange to understand whether someone was stating a finding, asking a question, repeating a rumor, or offering a hypothetical.
Access to that data and message trail is central. If it is missing, investigate connection and export options and help obtain the specific original thread, document, recording, or dataset. Say upfront that reconstructing the trail through an interview will be difficult and incomplete. My recollection may suggest search leads, but it cannot by itself establish provenance or the original source. Continue useful accessible branches and accurately mark inaccessible ones.
2. Follow the links backward without losing the variants
Build a source trail, keeping references sufficient to reopen each item. For each occurrence, record:
- Its actual wording and qualifiers, author or publisher where known, and publication or message date.
- The source it cites, links, attaches, or appears to paraphrase, distinguishing explicit attribution from your inference.
- Changes to scope, population, timeframe, certainty, quantities, or meaning relative to nearby versions.
Inspect linked originals and attachments. Follow citations through summaries to the document, data, interview, or other evidence behind them. Compare dates carefully: the earliest item found is not necessarily the origin, and a modified file date is not a publication date. Detect circular citations and repeated unsourced versions rather than counting them as independent confirmation. Use archived versions where accessible and needed, marking which version was examined.
When the trail branches, assign separate bounded agents to different citation paths, message threads, or archives where available. Give each the precise claim, known starting source, search scope, and a requirement to return actual inspected evidence with unresolved links. Let first searches finish before reconciling them. Keep a coverage record and use waves or compact fresh-session handoffs when the trail exceeds one context. Without subagents, investigate separate branches in manageable passes and describe their independence accurately.
3. Check the underlying evidence in context
Read enough of the strongest available underlying source to establish what it actually says. Examine definitions, methods, sample or population, dates, denominators, and limitations that affect this claim. If the claim rests on company data, inspect the relevant calculation, reporting period, and version where available. Recalculate a material number when the underlying data permits; report when it does not.
Have a separate verifier inspect the candidate source and compare it with the claim. Ask whether the source directly supports the wording, supports only a narrower statement, contradicts it, or cannot resolve it. Give the verifier the sources and claim, not an instruction to validate the lead's preferred conclusion. Resolve disagreements by revisiting evidence. If a separate verifier is unavailable, do a distinct source-verification pass after discovery, reopening the candidate evidence and comparing it with the exact claim. Use a fresh-session handoff where possible; disclose when verification shares discovery context. A source can be the actual origin of a claim without proving that claim true.
4. Deliver the finding and a usable correction
Lead with what the evidence supports, followed by the traceable path and material changes in retelling. Distinguish a verified original source, the earliest accessible occurrence, and an unresolved origin. Include direct links or usable source references and short quotations only where they help make a consequential difference clear.
Offer a corrected, properly qualified formulation I can use, preserving unresolved details rather than filling them in. If the trail runs out, state exactly where and what missing evidence would be needed to continue. A transparent unresolved finding is a valid investigation result; do not promote repeated claims or an interview reconstruction into verified provenance. Stop with the source account and proposed wording, without contacting message authors or changing shared documents.
Done when
You have a traceable account of the claim and the evidence behind it, including changes in retelling and any unresolved origin or missing links.
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.