Repurpose Longform Content Prompt

Use ChatGPT to transform the supplied long-form material into a review-ready set of channel-specific content assets.

Inputs
- Long-form material or transcript: [Longform source]
- Campaign objective: [Repurposing goal]
- Intended audience and relevant audience segments: [Target audience]
- Requested channels, asset formats, and quantity per format: [Channels, formats, and quantities]
- Brand tone, terminology, prohibited language, and representative examples: [Brand voice and examples]
- Length limits, deadlines, usage rights, privacy restrictions, regulated-claim rules, CTA restrictions, and required approvers: [Constraints, rights, and approval rules]
- Acceptance criteria and available performance benchmarks: [Success criteria]

Input and access rules
1. Treat the long-form source, audience, objective, and requested assets as required inputs. If the source is missing, inaccessible, or provided only as a URL that ChatGPT cannot inspect, request the text, transcript, or accessible file and stop before drafting.
2. If the audience, objective, or requested channels are ambiguous enough to change the message materially, ask only the questions needed to resolve that ambiguity. If a noncritical preference is absent, continue with a conservative assumption and record it.
3. Inspect only material present in the conversation or files ChatGPT can actually read. Do not imply that a linked page, recording, analytics account, publishing platform, or brand system was accessed when it was not.
4. Identify conflicting instructions or source statements. Preserve the conflict in the output instead of choosing a convenient version without evidence.

Evidence and editorial rules
- Build every factual claim, statistic, quotation, case-study result, date, product capability, and attributed opinion from the supplied source. Attach a source reference using a page, section, paragraph, speaker, or timestamp when available.
- Distinguish direct source facts, faithful paraphrases, creative framing, user-supplied requirements, assumptions, and unresolved claims. Creative hooks may reframe an idea but must not introduce a new factual promise.
- Reproduce direct quotations exactly. If wording is shortened, mark the omission and confirm that the meaning is unchanged. Otherwise use a clearly identified paraphrase.
- Do not fabricate statistics, testimonials, customer outcomes, credentials, urgency, scarcity, endorsements, citations, or performance predictions.
- Preserve qualifications and uncertainty from the source. Do not convert correlation into causation, an example into a universal rule, or an aspiration into a proven result.
- Avoid copying long passages when transformation or summary is sufficient. Flag uncertain ownership, licensing, attribution, confidentiality, personal data, medical or financial claims, and other material requiring legal, compliance, or subject-matter review.
- Do not expose private or confidential information. Redact unnecessary personal data and stop drafting affected assets if safe redaction would destroy essential meaning.

Repurposing workflow
1. Parse the source into a content inventory: central thesis, audience problem, supporting arguments, evidence, examples, memorable language, objections, caveats, and available calls to action.
2. Create source-reference IDs for reusable ideas and claims. Record where each item appears and whether it is safe to quote, paraphrase, or use only after approval.
3. Design a campaign arc rather than mechanically slicing the source. Assign each proposed asset a distinct communication job such as awareness, education, objection handling, proof, conversation, or conversion.
4. Match each asset to the conventions of its requested channel and format. Adjust hook density, pacing, line breaks, thread structure, script beats, subject line, preheader, CTA, visual direction, and accessibility fields where relevant. Apply only platform limits supplied in the inputs or otherwise known with confidence; label uncertain limits for confirmation.
5. Draft the requested quantity of assets. Keep the source meaning intact while varying the angle, opening, structure, and CTA. Do not present superficial truncations as distinct assets.
6. For each asset, identify the source-reference IDs supporting its substantive claims. Mark source-independent creative devices, such as a question hook, as creative framing rather than evidence.
7. Review the complete set for unsupported claims, duplicated angles, context loss, quote distortion, tone drift, channel mismatch, privacy exposure, rights concerns, inaccessible visual concepts, and conflicting CTAs.
8. Recommend a release sequence only when the objective and requested channels support one. Explain sequencing dependencies and identify assets blocked by unresolved claims or approvals.

Authority and completion boundaries
- Produce drafts, analysis, and review recommendations only. Do not claim to publish, schedule, send, approve, obtain permission for, or measure any asset.
- Treat legal, compliance, brand, subject-matter, and publication approvals as human decisions governed by the supplied approval rules.
- Label each asset Draft, Blocked, or Ready for human review. Never label an asset approved, published, tested, or performance-validated without explicit evidence that the corresponding action occurred.
- If requested content would materially misrepresent the source, violate stated rights, reveal protected information, or make an unsupported consequential claim, do not draft that content. Explain the blocking issue and propose a safer alternative.

Required output

1. Intake status
- State the objective, audience, requested asset inventory, supplied constraints, accessible materials, blocking omissions, bounded assumptions, and input conflicts.

2. Source evidence map
Provide a table with: source-reference ID; source location or timestamp; source idea or claim; evidence type; exact quote or faithful paraphrase; necessary qualification; rights or sensitivity note; permitted reuse mode; and confidence.

3. Campaign architecture
Provide a table with: asset ID; channel and format; communication job; audience segment or journey stage; angle; supporting source-reference IDs; CTA; planned relationship to other assets; and draft status.

4. Channel-ready draft pack
For every asset, provide:
- Asset ID, channel, format, and status
- Intended audience and communication job
- Primary copy or script in the requested structure
- One materially different hook or subject-line alternative
- CTA
- Source-reference IDs for factual content
- Visual or production brief when relevant
- Accessibility support such as alt-text direction, captions, transcript needs, or pronunciation notes when relevant
- Character, word, duration, or structural count when a supplied limit applies
- Required human reviews and unresolved notes

For threads, number each post and verify narrative continuity. For short-form video, separate spoken lines, on-screen text, shot or B-roll direction, and approximate timing. For email, separate subject line, preheader, body, and CTA. For other formats, use fields native to that format rather than forcing a generic layout.

5. Claims and editorial risk register
Provide a table with: asset ID; affected wording; risk category; source evidence; uncertainty or conflict; potential consequence; mitigation; required approver; and disposition as pass, revise, or blocked.

6. Verification and acceptance record
Evaluate the generated drafts with a table containing: check; expected condition; observed condition in the draft; evidence location; result as pass, fail, or blocked; and required correction.

At minimum, verify:
- Every substantive factual claim maps to a valid source-reference ID.
- Quotations match the source and retain material context.
- Caveats, uncertainty, and attribution survive repurposing.
- Requested channels, formats, and quantities reconcile with the delivered asset inventory.
- Each asset has a distinct angle and is not merely a shortened duplicate.
- Copy follows supplied voice, terminology, CTA, and format constraints.
- Applicable length or timing limits are calculated rather than guessed.
- Privacy, rights, regulated-claim, and approval requirements are addressed.
- Accessibility fields are included where the format requires them.
- No draft is described as published, approved, tested, or measured without execution evidence.

7. Approval and handoff queue
List each asset with its current state, owner or required reviewer if supplied, blocking issue, evidence needed to clear it, and the smallest safe next action. End by reconciling the number of requested assets against the number delivered, blocked, and awaiting clarification.