JP

Case study

One article in. Eleven platform-ready assets out.

The engine that ended the content treadmill — built for Unpromptable first, offered to clients now.

1 → 11

assets from every longform article

~3 hrs

of manual repurposing saved per cycle

Engagement: Internal build — offered to clients as a content engine

What does an AI content repurposing system actually look like?

One trigger, one parser, four parallel branches, one writing engine. A finished article enters a spreadsheet row and leaves as 11 platform-specific drafts, with the AI writing from examples of past high-performing content rather than a generic template. The original writing stays human; only the repetitive distribution work is automated.

Who they were

This one is mine. Unpromptable publishes across Substack, Medium, and LinkedIn — written by one person carrying a full client roster.

The problem

One deep article takes hours to write and lands on one platform. Repurposing it by hand for LinkedIn, Medium, and Substack Notes ate roughly four more hours per piece. The better client work went, the less I published — and the less visible the brand became. Success was feeding the plateau.

The writing itself was never a candidate for automation. That stays human. Repurposing is different: repetitive, formulaic, and exactly the kind of work a system should absorb.

What we built

  1. 01

    Trigger

    Drop a finished article into a Google Sheet row. That's the entire manual step.

  2. 02

    Parser

    Extraction and cleaning nodes strip the piece down to its working material.

  3. 03

    Four parallel branches

    Medium articles, a LinkedIn carousel, standalone LinkedIn posts, and Substack Notes — each branch shaped for its platform, not just shortened.

  4. 04

    Writing engine

    Claude via OpenRouter, loaded with high-performing past examples per branch so the output matches the voice and the format it's headed for.

  5. 05

    Output

    Eleven pre-formatted drafts land back in the spreadsheet, ready for a human pass and scheduling.

What changed

Every longform piece now fans out into 11 distribution-ready assets: 2 Medium articles, 1 LinkedIn carousel, 4 LinkedIn posts, and 4 Substack Notes. Around three hours of manual repurposing disappear from every cycle.

The effort compounds instead of resetting. One writing session feeds a week of distribution — a flywheel where there used to be a treadmill.

Automate the repetitive. Never the voice. The original thinking stays unpromptable — the distribution scales.

Read the full build story on Substack

More case studies

Your version of this

Every one of these started as a conversation.

Tell me what you're sitting on — expertise, archives, a content bottleneck — and I'll tell you what I'd build.

Not ready to talk yet? Start with the AI Readiness Diagnostic.