Turn one video transcript into five content pieces
You are a content marketing specialist for {{COMPANY_NAME}}, {{COMPANY_DESCRIPTION}}. Your task is to analyze a video transcript and create five pieces of optimized content from it. I will provide the transcript below.
ANALYZE THE TRANSCRIPT FOR
1. The main topic and core message.
2. Key insights, data points, or takeaways.
3. Names of speakers or presenters — if none are given, refer to them as "{{COMPANY_NAME}}'s experts."
4. Any unique selling points or genuinely valuable information.
CREATE THE FOLLOWING FIVE DELIVERABLES
1. YouTube title (60-70 characters) — clear, attention-grabbing, optimized for search, with a hook that promises value without misrepresenting the content.
2. YouTube description (150-300 words) — summarize the key points, work in relevant keywords naturally, structure as an engaging intro, main points, and a strong call to action.
3. YouTube tags (up to 400 characters) — relevant tags drawn from the transcript, including at least 5 multi-word tag phrases and the keywords used in the title and description.
4. YouTube thumbnail hook (up to 60 characters) — optimized for click-through, resonant with the audience, and an accurate reflection of the video's actual content.
5. LinkedIn post (1300-1500 characters) — written from {{COMPANY_NAME}}'s perspective, highlighting the most valuable insight from the video, closing with 3-5 relevant hashtags, structured for easy reading with short paragraphs.
FOR ALL FIVE DELIVERABLES
- Write in third person.
- Use simple, direct language and short sentences.
- Avoid AI-giveaway phrases and marketing hype — say "here's how it works," not "let's dive into this game-changing solution."
- Be concise; cut anything that doesn't earn its place.
- Write in a natural, conversational tone. Starting a sentence with "and" or "but" is fine.
- Skip promotional words like "revolutionary" or "transformative" — describe what it actually does.
- Cut unnecessary adjectives and adverbs.
OUTPUT FORMAT
<youtube_title>[title]</youtube_title>
<youtube_description>[description]</youtube_description>
<youtube_tags>[tags]</youtube_tags>
<youtube_thumbnail_hook>[hook]</youtube_thumbnail_hook>
<linkedin_post>[post]</linkedin_post>
TRANSCRIPT
{{VIDEO_TRANSCRIPT}}One transcript in, five ready-to-publish assets out — a YouTube title, description, tags, thumbnail hook, and a LinkedIn post — all pulled from the same read of the source material so they don’t drift into inconsistent claims about what the video actually covers. It earns its keep the moment you’re repurposing video content regularly enough that writing each piece by hand has become the bottleneck.
What to fill in
{{COMPANY_NAME}}— the company whose voice this represents; used throughout for the third-person framing.{{COMPANY_DESCRIPTION}}— one line on what the company does. Keep it short; this is context, not a full brand brief.{{VIDEO_TRANSCRIPT}}— the transcript itself. Clean, speaker-labeled transcripts produce noticeably better output than raw auto-captions, since the model can actually attribute specific points to specific speakers.
Why it works
Doing the analysis once and generating all five deliverables from it, rather than prompting separately for each, keeps the core message consistent across formats. A common failure mode with channel-by-channel prompting is that the YouTube description emphasizes one takeaway and the LinkedIn post emphasizes a different one, so anyone who sees both gets a slightly different story. A single analysis pass removes that risk by construction.
The explicit style rules matter as much as the structure. Character and word limits alone don’t stop a model from writing in a generic, over-enthusiastic marketing voice — telling it directly to avoid hype words and AI-tell phrases, and showing a couple of before/after examples, does far more to keep the output usable without a heavy edit.
Where it falls short
It doesn’t know your actual brand voice beyond a one-line description, so if your company has a distinctive tone, expect to edit rather than publish directly. It also assumes the video has one clear core message — feed it a rambling multi-topic panel discussion and the five deliverables will read as competently written but strangely unfocused, because there was never one thing to focus on.
Free skills that do part of this job permanently, so you stop pasting the instruction every time.