AI Agents for Content Repurposing

Turn one blog post into LinkedIn posts, email newsletters, and tweets automatically. New content every Tuesday from work you've already done.

01the problem

Your agency produces great long-form content but it dies after one publish. A blog post goes up, gets shared once, and that's it. Turning one piece into LinkedIn posts, email copy, and social snippets takes another 2-3 hours. So it doesn't happen. Your best thinking reaches a fraction of the audience it could.

02what the agent does

Every Tuesday at 9 AM, the agent checks for any new long-form content from the past week (blog posts, case studies, newsletters in Notion or your content folder). For each piece, it generates: 3 LinkedIn post variants (hook-driven, story-driven, insight-driven), an email newsletter draft, and 5 tweets. Everything matches the original tone. You review and schedule.

Example output

agent output

LinkedIn Post — Hook-driven

Every agency has the same margin killer and most don't catch it until the project is underwater.

It's not bad clients. It's scope creep that nobody tracks.

We built an AI agent that runs every two weeks. It compares the original SOW against every task completed, every Slack request, every verbal agreement from calls.

Last month it caught 13 hours of unbilled work across 3 clients.

The fix wasn't billing more. It was having the conversation before the work piled up.

What's your system for catching scope drift?


Email Newsletter (150 words)

Subject: The 13 hours nobody noticed

Every agency leaks hours. Not through bad work. Through work that grows past what was agreed.

We set up an AI agent that checks scope every two weeks...

Try it yourself

Copy these prompts into Claude Cowork. Connect your tools. The agent handles the rest.

01
Repurpose a single piece of content
Point the agent at any article or document and get distribution-ready content back.
Read [ARTICLE URL or FILE PATH]. Create the following, matching the original tone exactly:

1. LinkedIn post (hook-driven): Open with a counterintuitive claim. 3-5 short paragraphs. End with a question. Max 1,200 characters.
2. LinkedIn post (story-driven): Open with a specific moment. Build to the insight. Max 1,200 characters.
3. LinkedIn post (insight-driven): Open with a data point. 2-3 supporting points. Close with one implication. Max 1,200 characters.
4. Email newsletter: 150-250 words. Conversational. One CTA.
5. 5 tweets: Each under 280 characters. One idea per tweet. No hashtags.

FAQ

Common questions about setting up this agent.

Will the repurposed content sound like AI slop?+
Only if the original does. The agent is instructed to match the tone and voice of the source material exactly. If your blog post has a specific point of view and real examples, the LinkedIn posts will too. The quality of the output mirrors the quality of the input.
Can this work for client content, not just our own?+
Yes. Agencies that produce content for clients can run this for each client. Store each client's brand voice and guidelines in their project folder. The agent reads those before generating. Each client gets repurposed content that sounds like them, not like a generic AI.
What if we publish on WordPress, not Notion?+
The agent can read from any connected source or from files in your Cowork folder. If you publish on WordPress, you can point it to the URL of the published post. It reads the content and repurposes from there. The source format doesn't matter.

Want this running for your team?

I set up AI agents like this for agencies. 30 minutes, no strings.