AI Social Media Growth
0/15 complete

Module 03 · Content Production With AI

Repurposing One Video Into Five Pieces of Content

Repurposing means extracting several useful ideas from one verified source video, then rebuilding each idea for the way people consume content at its destination. It is not uploading the same crop and caption to five feeds.

By the end, you will have a source-of-truth brief and a complete five-asset pack: one short clip, one quote graphic, one text post, one carousel, and one thread. Each asset will keep the source video's meaning while using a different opening, format, and action.

// concept

Build a Source-of-Truth Brief

Start with a video you own or are authorized to use. Transcribe it, add timestamps, and correct names, prices, Urdu words, and product details by replaying the recording.

Create this brief before asking AI to write anything:

FieldWhat to record
SourceFile name, owner, recording date, and permission
AudienceOne specific viewer group
Core messageOne sentence the video actually supports
Verified pointsTimestamped claims, demonstrations, or exact quotations
ExclusionsPrivate details, mistakes, or off-topic sections not to reuse
Desired actionOne sensible next step, not a performance promise

Use <source> tags around the verified transcript. If the source never mentions a discount, result, or customer reaction, the derivatives must not invent one.

// concept

Extract Five Decisions, Not Five Copies

Give each asset a different job: a clip demonstrates, a quote graphic freezes one line, a text post explains a decision, a carousel teaches a sequence, and a thread adds commentary.

Asset and destinationNative openingFormat choiceClosing action
TikTok clipSpoken problem in the first beat9:16, 15–30 seconds as an editorial target, on-screen captionsAsk viewers to save the packing checklist
Instagram quote graphicSix-to-twelve-word source quotation1080 × 1080 PNG; short caption adds contextAsk one answerable question
LinkedIn text postA business decision, not a video teaserShort paragraphs and a readable first lineInvite peers to share their process
LinkedIn document carouselA useful outcome on slide 1One PDF with six consistent 4:5 pagesTell readers what the final slide helps them check
Threads threadA direct observationFour short connected posts; check the current composer limitAsk which step deserves a demonstration

These are working choices, not claims about reach. Platform support changes, so confirm current upload requirements in Sources. Meta currently documents Reel support up to 9:16; preview the cover and captions inside the app.

// concept

Rebuild the Opening, Body, and Action

For each derivative, write three lines before designing:

  1. Opening: why someone at this destination should stop.
  2. Body: the one source-backed point this format can carry well.
  3. Action: what the viewer can reasonably do next.

Then use this complete prompt. Replace the sample fields and transcript; do not paste customer phone numbers, addresses, order screenshots, or private client material.

// prompt — copy me22 lines
You are a content repurposing editor. Use only the verified material inside
<source>. Do not add facts, prices, results, quotations, or promises.

Audience: Pakistani home-business owners who pack fragile products.
Core message: A repeatable packing order reduces avoidable movement in the box.
Voice: practical, calm, English with Roman Urdu only if it appears in the source.

Create exactly five distinct assets:
1) TikTok: a 20–25 second 9:16 clip plan with timestamped source cuts,
   on-screen text, caption, and one save action.
2) Instagram: a square quote graphic using one exact source sentence,
   plus a two-sentence caption and one question.
3) LinkedIn: a 120–180 word text post about the process decision.
4) LinkedIn: a six-slide PDF carousel; one headline and <=25 words per slide.
5) Threads: a four-post thread; each post must add a different idea.

For every asset, show: source timestamps, opening, body, action, and any claim
that needs human verification. Make the openings and actions different.

<source>
[Paste the corrected, timestamped transcript here.]
</source>

Trace every sentence to a timestamp and delete unsupported additions. Preview each export in its destination composer because text-safe areas and crop previews can change.

// worked_example

Worked Example

This is a hypothetical sample, not a real result. A Karachi home baker owns a 60-second video called eid-box-packing.mp4. Its source says: 00:00–00:08 “A pretty box still fails if the cookies can move.” 00:09–00:22 fold the insert; 00:23–00:36 separate layers; 00:37–00:49 close empty space; 00:50–01:00 test the closed box and inspect movement.

The prompt above produces this five-asset pack:

  1. TikTok clip: cuts 00:00–00:08, 00:23–00:36, and 00:50–00:58. Opening text: “Packing check: can the cookies move?” Body labels: “separate layers” and “test before dispatch.” Caption: “Sample packing sequence from one owned video. Save it for your next packing day.”
  2. Instagram quote graphic: “A pretty box still fails if the cookies can move.” The caption identifies it as a sample packing principle and asks, “What part of your box moves first?”
  3. LinkedIn text post: “Our packaging review starts with movement, not decoration.” It explains the insert, layer separation, and gentle closed-box check, then asks other small sellers which step they document for staff.
  4. LinkedIn carousel: slide 1 “Can the product move?”; slide 2 “Fold the insert”; slide 3 “Separate each layer”; slide 4 “Close empty space”; slide 5 “Test the closed box”; slide 6 “Reopen and inspect.” The post description says it is a sample checklist derived from the video.
  5. Threads thread: (1) “A neat box is not yet a checked box.” (2) “This sample process starts by limiting empty space.” (3) “Layers are separated before the lid closes.” (4) “The final step is a gentle movement check, then inspection. Which step should become a full demo?”

The bad draft resized the 60-second export, opened every asset with “Watch how we pack,” and invented “damage-free delivery.” The correction removes that claim and assigns one source-backed idea to each format.

// failure_cases

Failure Cases to Diagnose

6 cases to diagnose

  • Resize-only repurposing

    all five assets have the same opening and sequence. Fix it by assigning a separate job and source segment to each asset.

  • Transcript drift

    the graphic quotation is cleaner than what the speaker said. Either quote exactly or label a faithful paraphrase; never use quotation marks around rewritten speech.

  • Unsupported commercial claim

    AI adds “best seller,” delivery coverage, or a result absent from the video. Delete it and check every remaining claim against timestamps.

  • Contextless clip

    the cut starts with “that is why” or refers to an unseen item. Add a source-backed setup or choose a self-contained cut.

  • Unreadable mobile export

    captions sit behind interface controls or carousel text requires zooming. Preview in the actual composer and shorten the copy.

  • Identical actions

    every derivative says “buy now,” regardless of purpose. Match the action to the asset: save, answer, compare, or request the full demonstration.

// pakistan_angle

Pakistan Angle

For an Urdu or Roman-Urdu source, preserve the speaker's actual wording instead of translating everything into polished English. Check automatic captions manually for product names and similar-sounding Urdu words. A bilingual quote graphic can work, but write both versions yourself and ensure they carry the same meaning; AI transliteration is not the source of truth.

Load-shedding and mobile-data limits favour a staged workflow: download or transcribe the source once, produce lightweight text drafts offline, and queue high-resolution exports when electricity and connectivity are stable. Keep a compressed review copy for WhatsApp approval, but publish from the final master. Redact phone numbers, customer addresses, JazzCash/easypaisa receipts, and order labels before uploading footage or transcripts to any AI service.

// hands_on

Hands-On Exercise

6 steps

  1. Choose one 45–90 second video you own and record its file name, permission, audience, and intended action.

  2. Produce a corrected timestamped transcript and a source-of-truth brief using the table above.

  3. Run the transcript-to-assets prompt, then verify every output line against a timestamp.

  4. Finish one clip, quote graphic, text post, carousel, and thread. Give each a distinct opening, format, and action.

  5. Preview the five assets in their destination composers; record any crop, caption, or page-order fix.

  6. Save the pack with the source brief and a simple source map. Done means all five native assets exist and every factual line can be traced to the owned video.

// completion_rubric

Completion Rubric

5 checks — tick as you verify

0/5

// sources

Sources

// check_yourself

Check yourself

4 questions · answers and options are taken word-for-word from this course

0/4
  1. 1 / 4 · diagnose

    Your work shows this failure mode: “Resize-only repurposing.” The lesson describes it like this: “All five assets have the same opening and sequence.” What does the lesson tell you to do about it?