Use this before define acceptance criteria when the notes are rough and ChatGPT should ask clarifying questions first.
Act as a careful assistant for Product Managers.
Task: help me define acceptance criteria. Target result: acceptance criteria.
Source material I can provide: [source_material]. Typical source for this task is feature goal, edge cases, roles, data states, and failure behavior.
Audience or stakeholder: [audience]. The output must work for a product team, stakeholder, customer researcher, or release owner.
Task-specific focus: given-when-then states, edge cases, and testable completion.
Goal: [goal]. Constraints: [constraints]. Do not add facts that are not in the source material.
Run mode: Run this as intake: ask the questions needed before writing, then wait for answers if the source material is missing.
Stop rule: Stop before creating the final asset if the audience, source material, or review owner is unclear.
Return a question list grouped by audience, source material, constraints, and review owner.
Before the answer, ask up to 3 clarifying questions if the source material is too thin.
After the answer, include a human review section focused on acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence; verify source notes, examples, constraints, and reviewer judgment; and respect this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Check cue: The user should leave with a short context pack and a safe next prompt, not a finished answer.
- [source_material]
- The notes, facts, examples, or raw material behind define acceptance criteria.Example: feature goal, edge cases, roles, data states, and failure behavior
- [audience]
- Who will read, use, approve, or act on the output.Example: a product team, stakeholder, customer researcher, or release owner
- [goal]
- The decision or work outcome the response should support.Example: make acceptance criteria easier to review, adapt, and use in a real product managers workflow
- [constraints]
- Rules, tone, length, channel, privacy limits, and required sections.Example: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
- [review_lens]
- The most important thing a human should check after the answer.Example: acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence
- [task_focus]
- The task-specific detail that keeps this prompt from becoming generic.Example: given-when-then states, edge cases, and testable completion
Expected output
Expect a question list grouped by audience, source material, constraints, and review owner that explicitly separates source-based content from assumptions and ends with a review pass for acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Follow-up prompt
Now improve this working version into acceptance criteria by tightening acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence, emphasizing given-when-then states, edge cases, and testable completion, removing unsupported claims, and giving me one stronger version for a product team, stakeholder, customer researcher, or release owner.
Human review
Check whether the answer uses only provided context, handles source notes, examples, constraints, and reviewer judgment, fits a product team, stakeholder, customer researcher, or release owner, reflects given-when-then states, edge cases, and testable completion, and respects this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Best for: Starting define acceptance criteria when the source material still needs shape. Use when: Use before asking ChatGPT for define acceptance criteria so the model has enough task-specific context.
Use this when the source material is ready and the answer needs to become acceptance criteria.
Act as a careful assistant for Product Managers.
Task: help me define acceptance criteria. Target result: acceptance criteria.
Source material I can provide: [source_material]. Typical source for this task is feature goal, edge cases, roles, data states, and failure behavior.
Audience or stakeholder: [audience]. The output must work for a product team, stakeholder, customer researcher, or release owner.
Task-specific focus: given-when-then states, edge cases, and testable completion.
Goal: [goal]. Constraints: [constraints]. Do not add facts that are not in the source material.
Run mode: Run this as the first usable version: use the supplied fields, label assumptions, and produce the main artifact.
Stop rule: Stop if the request asks you to invent facts, evidence, credentials, numbers, or private details.
Return a scoring table with levels, observable evidence, and reviewer notes.
Before the answer, ask up to 3 clarifying questions if the source material is too thin.
After the answer, include a human review section focused on acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence; verify source notes, examples, constraints, and reviewer judgment; and respect this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Check cue: The user should get a working version they can inspect against the supplied notes.
- [source_material]
- The notes, facts, examples, or raw material behind define acceptance criteria.Example: feature goal, edge cases, roles, data states, and failure behavior
- [audience]
- Who will read, use, approve, or act on the output.Example: a product team, stakeholder, customer researcher, or release owner
- [goal]
- The decision or work outcome the response should support.Example: make acceptance criteria easier to review, adapt, and use in a real product managers workflow
- [constraints]
- Rules, tone, length, channel, privacy limits, and required sections.Example: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
- [review_lens]
- The most important thing a human should check after the answer.Example: acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence
- [task_focus]
- The task-specific detail that keeps this prompt from becoming generic.Example: given-when-then states, edge cases, and testable completion
Expected output
Expect a scoring table with levels, observable evidence, and reviewer notes that explicitly separates source-based content from assumptions and ends with a review pass for acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Follow-up prompt
Now improve this working version into acceptance criteria by tightening acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence, emphasizing given-when-then states, edge cases, and testable completion, removing unsupported claims, and giving me one stronger version for a product team, stakeholder, customer researcher, or release owner.
Human review
Check whether the answer uses only provided context, handles source notes, examples, constraints, and reviewer judgment, fits a product team, stakeholder, customer researcher, or release owner, reflects given-when-then states, edge cases, and testable completion, and respects this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Best for: Turning prepared context into acceptance criteria. Use when: Use before asking ChatGPT for define acceptance criteria so the model has enough task-specific context.
Use this when define acceptance criteria repeats often enough to become acceptance criteria prompt pattern with source notes, constraints, and review checklist.
Act as a careful assistant for Product Managers.
Task: help me define acceptance criteria. Target result: acceptance criteria.
Source material I can provide: [source_material]. Typical source for this task is feature goal, edge cases, roles, data states, and failure behavior.
Audience or stakeholder: [audience]. The output must work for a product team, stakeholder, customer researcher, or release owner.
Task-specific focus: given-when-then states, edge cases, and testable completion.
Goal: [goal]. Constraints: [constraints]. Do not add facts that are not in the source material.
Run mode: Run this as a repeatable workflow: separate one-time facts from fields that should change next time.
Stop rule: Stop if the reusable version would preserve private details or hide a human approval step.
Return a reusable step-by-step workflow with inputs, checks, and follow-up prompts.
Before the answer, ask up to 3 clarifying questions if the source material is too thin.
After the answer, include a human review section focused on acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence; verify source notes, examples, constraints, and reviewer judgment; and respect this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Check cue: The user should get reusable fields, a run order, and a reject-if rule for the next use.
- [source_material]
- The notes, facts, examples, or raw material behind define acceptance criteria.Example: feature goal, edge cases, roles, data states, and failure behavior
- [audience]
- Who will read, use, approve, or act on the output.Example: a product team, stakeholder, customer researcher, or release owner
- [goal]
- The decision or work outcome the response should support.Example: make acceptance criteria easier to review, adapt, and use in a real product managers workflow
- [constraints]
- Rules, tone, length, channel, privacy limits, and required sections.Example: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
- [review_lens]
- The most important thing a human should check after the answer.Example: acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence
- [task_focus]
- The task-specific detail that keeps this prompt from becoming generic.Example: given-when-then states, edge cases, and testable completion
Expected output
Expect a reusable step-by-step workflow with inputs, checks, and follow-up prompts that explicitly separates source-based content from assumptions and ends with a review pass for acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Follow-up prompt
Now improve this working version into acceptance criteria by tightening acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence, emphasizing given-when-then states, edge cases, and testable completion, removing unsupported claims, and giving me one stronger version for a product team, stakeholder, customer researcher, or release owner.
Human review
Check whether the answer uses only provided context, handles source notes, examples, constraints, and reviewer judgment, fits a product team, stakeholder, customer researcher, or release owner, reflects given-when-then states, edge cases, and testable completion, and respects this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Best for: Creating a reusable process for repeated define acceptance criteria work. Use when: Use when define acceptance criteria repeats often enough to need a standard process.
Use this after there is already working copy and the main need is acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Act as a careful assistant for Product Managers.
Task: help me define acceptance criteria. Target result: acceptance criteria.
Source material I can provide: [source_material]. Typical source for this task is feature goal, edge cases, roles, data states, and failure behavior.
Audience or stakeholder: [audience]. The output must work for a product team, stakeholder, customer researcher, or release owner.
Task-specific focus: given-when-then states, edge cases, and testable completion.
Goal: [goal]. Constraints: [constraints]. Do not add facts that are not in the source material.
Run mode: Run this as a review of existing copy: score the answer, name the weak sections, and propose repairs.
Stop rule: Stop if the copy cannot be traced back to the supplied source material or the reviewer is not named.
Return a scored review table with issues, fixes, and what still needs human judgment.
Before the answer, ask up to 3 clarifying questions if the source material is too thin.
After the answer, include a human review section focused on acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence; verify source notes, examples, constraints, and reviewer judgment; and respect this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Check cue: The user should get a decision about accept, repair, or reject before polishing the wording.
- [source_material]
- The notes, facts, examples, or raw material behind define acceptance criteria.Example: feature goal, edge cases, roles, data states, and failure behavior
- [audience]
- Who will read, use, approve, or act on the output.Example: a product team, stakeholder, customer researcher, or release owner
- [goal]
- The decision or work outcome the response should support.Example: make acceptance criteria easier to review, adapt, and use in a real product managers workflow
- [constraints]
- Rules, tone, length, channel, privacy limits, and required sections.Example: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
- [review_lens]
- The most important thing a human should check after the answer.Example: acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence
- [task_focus]
- The task-specific detail that keeps this prompt from becoming generic.Example: given-when-then states, edge cases, and testable completion
Expected output
Expect a scored review table with issues, fixes, and what still needs human judgment that explicitly separates source-based content from assumptions and ends with a review pass for acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Follow-up prompt
Now improve this working version into acceptance criteria by tightening acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence, emphasizing given-when-then states, edge cases, and testable completion, removing unsupported claims, and giving me one stronger version for a product team, stakeholder, customer researcher, or release owner.
Human review
Check whether the answer uses only provided context, handles source notes, examples, constraints, and reviewer judgment, fits a product team, stakeholder, customer researcher, or release owner, reflects given-when-then states, edge cases, and testable completion, and respects this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Best for: Finding weak spots in existing working copy. Use when: Use after product managers already have working copy and need to check acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Use this when the substance is right but the output needs to fit a table, checklist, email, outline, or script.
Act as a careful assistant for Product Managers.
Task: help me define acceptance criteria. Target result: acceptance criteria.
Source material I can provide: [source_material]. Typical source for this task is feature goal, edge cases, roles, data states, and failure behavior.
Audience or stakeholder: [audience]. The output must work for a product team, stakeholder, customer researcher, or release owner.
Task-specific focus: given-when-then states, edge cases, and testable completion.
Goal: [goal]. Constraints: [constraints]. Do not add facts that are not in the source material.
Run mode: Run this as format conversion: preserve the facts and change only the structure, order, or channel fit.
Stop rule: Stop if the requested format would require adding facts that were not in the original answer.
Return the same content reshaped without adding new facts.
Before the answer, ask up to 3 clarifying questions if the source material is too thin.
After the answer, include a human review section focused on acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence; verify source notes, examples, constraints, and reviewer judgment; and respect this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Check cue: The user should get a reshaped version plus a note showing what stayed unchanged.
- [source_material]
- The notes, facts, examples, or raw material behind define acceptance criteria.Example: feature goal, edge cases, roles, data states, and failure behavior
- [audience]
- Who will read, use, approve, or act on the output.Example: a product team, stakeholder, customer researcher, or release owner
- [goal]
- The decision or work outcome the response should support.Example: make acceptance criteria easier to review, adapt, and use in a real product managers workflow
- [constraints]
- Rules, tone, length, channel, privacy limits, and required sections.Example: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
- [review_lens]
- The most important thing a human should check after the answer.Example: acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence
- [task_focus]
- The task-specific detail that keeps this prompt from becoming generic.Example: given-when-then states, edge cases, and testable completion
Expected output
Expect the same content reshaped without adding new facts that explicitly separates source-based content from assumptions and ends with a review pass for acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Follow-up prompt
Now improve this working version into acceptance criteria by tightening acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence, emphasizing given-when-then states, edge cases, and testable completion, removing unsupported claims, and giving me one stronger version for a product team, stakeholder, customer researcher, or release owner.
Human review
Check whether the answer uses only provided context, handles source notes, examples, constraints, and reviewer judgment, fits a product team, stakeholder, customer researcher, or release owner, reflects given-when-then states, edge cases, and testable completion, and respects this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Best for: Changing the output format without changing the facts. Use when: Use when the answer needs a precise structure before product managers can review it.
Use this when the source material contains private, sensitive, or account-specific details.
Act as a careful assistant for Product Managers.
Task: help me define acceptance criteria. Target result: acceptance criteria.
Source material I can provide: [source_material]. Typical source for this task is feature goal, edge cases, roles, data states, and failure behavior.
Audience or stakeholder: [audience]. The output must work for a product team, stakeholder, customer researcher, or release owner.
Task-specific focus: given-when-then states, edge cases, and testable completion.
Goal: [goal]. Constraints: [constraints]. Do not add facts that are not in the source material.
Run mode: Run this as a sanitizing pass: replace private details with role-safe descriptions before writing.
Stop rule: Stop if names, identifiers, account details, confidential strategy, or one-time records are still present.
Return a sanitized prompt-ready summary plus a list of removed details.
Before the answer, ask up to 3 clarifying questions if the source material is too thin.
After the answer, include a human review section focused on acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence; verify source notes, examples, constraints, and reviewer judgment; and respect this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Check cue: The user should get a safe summary, removed-detail list, and a reusable version without sensitive data.
- [source_material]
- The notes, facts, examples, or raw material behind define acceptance criteria.Example: feature goal, edge cases, roles, data states, and failure behavior
- [audience]
- Who will read, use, approve, or act on the output.Example: a product team, stakeholder, customer researcher, or release owner
- [goal]
- The decision or work outcome the response should support.Example: make acceptance criteria easier to review, adapt, and use in a real product managers workflow
- [constraints]
- Rules, tone, length, channel, privacy limits, and required sections.Example: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
- [review_lens]
- The most important thing a human should check after the answer.Example: acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence
- [task_focus]
- The task-specific detail that keeps this prompt from becoming generic.Example: given-when-then states, edge cases, and testable completion
Expected output
Expect a sanitized prompt-ready summary plus a list of removed details that explicitly separates source-based content from assumptions and ends with a review pass for acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Follow-up prompt
Now improve this working version into acceptance criteria by tightening acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence, emphasizing given-when-then states, edge cases, and testable completion, removing unsupported claims, and giving me one stronger version for a product team, stakeholder, customer researcher, or release owner.
Human review
Check whether the answer uses only provided context, handles source notes, examples, constraints, and reviewer judgment, fits a product team, stakeholder, customer researcher, or release owner, reflects given-when-then states, edge cases, and testable completion, and respects this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Best for: Sanitizing context before asking ChatGPT for help. Use when: Use before adding sensitive context so private details stay out.
Use this for a quick pass when the user only needs the next few decisions for define acceptance criteria.
Act as a careful assistant for Product Managers.
Task: help me define acceptance criteria. Target result: acceptance criteria.
Source material I can provide: [source_material]. Typical source for this task is feature goal, edge cases, roles, data states, and failure behavior.
Audience or stakeholder: [audience]. The output must work for a product team, stakeholder, customer researcher, or release owner.
Task-specific focus: given-when-then states, edge cases, and testable completion.
Goal: [goal]. Constraints: [constraints]. Do not add facts that are not in the source material.
Run mode: Run this as a fast decision pass: give only the next actions, the missing input, and the main risk.
Stop rule: Stop if the user needs a full artifact, a legal answer, a policy decision, or unsupported factual claims.
Return a concise checklist with the next action and the main risk.
Before the answer, ask up to 3 clarifying questions if the source material is too thin.
After the answer, include a human review section focused on acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence; verify source notes, examples, constraints, and reviewer judgment; and respect this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Check cue: The user should get a narrow next step they can complete before opening a longer prompt.
- [source_material]
- The notes, facts, examples, or raw material behind define acceptance criteria.Example: feature goal, edge cases, roles, data states, and failure behavior
- [audience]
- Who will read, use, approve, or act on the output.Example: a product team, stakeholder, customer researcher, or release owner
- [goal]
- The decision or work outcome the response should support.Example: make acceptance criteria easier to review, adapt, and use in a real product managers workflow
- [constraints]
- Rules, tone, length, channel, privacy limits, and required sections.Example: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
- [review_lens]
- The most important thing a human should check after the answer.Example: acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence
- [task_focus]
- The task-specific detail that keeps this prompt from becoming generic.Example: given-when-then states, edge cases, and testable completion
Expected output
Expect a concise checklist with the next action and the main risk that explicitly separates source-based content from assumptions and ends with a review pass for acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence.
Follow-up prompt
Now improve this working version into acceptance criteria by tightening acceptance criteria quality, given-when-then states and edge cases, and decision-ready evidence, emphasizing given-when-then states, edge cases, and testable completion, removing unsupported claims, and giving me one stronger version for a product team, stakeholder, customer researcher, or release owner.
Human review
Check whether the answer uses only provided context, handles source notes, examples, constraints, and reviewer judgment, fits a product team, stakeholder, customer researcher, or release owner, reflects given-when-then states, edge cases, and testable completion, and respects this boundary: Prompts should surface assumptions and evidence gaps instead of pretending strategy is decided.
Best for: Getting a quick decision checklist before spending more time. Use when: Use when time is short and the user needs the next action, not a full answer.