← Back to library

Turn a recorded customer interview into three drafts

Wilbert Kramer, a communications lead who films customer stories, built a Cowork skill to stop losing momentum after each shoot. It reads a folder of video, briefing notes and photos, transcribes, then drafts blog, LinkedIn and newsletter pieces behind gates between formats.

  • OutcomeGet three pieces of copy, long article first, then a social post and a short email section, out of one customer video, approving each before the next is written.
  • SetupUnder an hour
  • LevelIntermediate
  • CreditWilbert Kramer
View original
Works with
Claude CoworkClaude Code

Published Oct 23, 2026Updated Oct 23, 2026AI dry run on sample input: passed Oct 23, 2026Source verified Oct 24, 2026

When should you use it?

  • You record customer or partner interviews
  • Repurposing stalls after the edit
  • You want to steer every piece rather than accept a batch

Not for: Publishing without the customer's approval of quotes; Interviews you do not have rights to reuse

What do you need, and what do you get?

You need

  • Folder with the video, a briefing document and photos
  • Style notes for each channel

You get

  • Article draft
  • LinkedIn draft
  • Newsletter paragraph
  • All three staged as drafts in their tools

How does it work, step by step?

  1. Standardize the input folderDecide what every interview folder contains so the skill always knows where to look.
  2. TranscribeUse a local speech-to-text tool or a browser one and save the text alongside the video.
  3. Draft the article firstWrite the long piece in a reporting style and stop for your review.
  4. Revise with specific feedbackGive pointed notes such as more customer quotes or a shorter intro until you approve.
  5. Derive the short piecesOnly after approval, write the LinkedIn and newsletter versions, then save all three as drafts.

The prompt

The original prompt by Wilbert Kramer, word for word, with a link to where it was published. Below it is our adapted version with fill-in blanks, if you want it. Works in Claude Cowork and Claude Code. Values in lime are placeholders.

Original prompt by Wilbert Kramer

Word for wordby Wilbert Kramer
I want to create a Cowork skill that turns a client interview (video + briefing) into three content outputs:
1. A blog post (client case study) → [your CMS] draft
2. A LinkedIn post → LinkedIn draft
3. A newsletter paragraph → [your email tool] draft Input: a folder with brief.txt, video.mp4, and photos. The flow should work step-by-step with human review after every generation step. Nothing gets published automatically. My style for case studies is [describe your style here — journalistic/narrative/promotional/etc.]. The LinkedIn post is on behalf of [name + title]. Create the SKILL.md and a references/prompts.md, and package it as a .skill file.

Exactly as published by Wilbert Kramer in Content Repurposing In Claude Cowork.

Our adapted version (optional)

Adapted prompt0 variables
Read the interview folder I point you to (video, briefing notes, photos). Transcribe the video and save the transcript. Write a reported-style article that quotes the speaker accurately, then stop and wait for my feedback. Revise until I say approved. Only then derive a social post and an email blurb from the approved article, each waiting for approval. Finally save all three as drafts; never publish.

What can go wrong?

  • Skipping the review gates lets one weak article spread its problems into every channel.
  • Check quotes against the transcript; summarizing can change meaning.

Our take

The design choice we like is the gate between pieces. Most repurposing setups spit out every format at once and you edit three flawed drafts; this one makes you approve the spine first. Wilbert's skill starts by standardizing what each interview folder holds, which means the video, briefing notes and photos are always where the skill expects them. Transcribe locally if you can, so less raw footage leaves your machine. The long article comes first in a reporting style and stops for review; give specific notes, such as more customer quotes or a shorter intro, until you approve it. Only then are the LinkedIn and newsletter versions written, so one weak draft does not spread its problems into every channel. Check every quote against the transcript, since summarizing can change meaning, and get the customer's approval before anything is published. All three end up as drafts, which keeps publishing in human hands.

Where the time goes back: The transcription and three rewrites that used to drain momentum after a shoot become a guided review session.

FAQ

Why one step at a time?
The author wanted to steer after each output; nothing advances until he explicitly approves.
Does it post automatically?
No. Everything stays unpublished until a person releases it.

Source and credit

The original prompt on this page is by Wilbert Kramer, shown word for word with a link to where it was published. The summary, notes and adapted version are ours.

Original
Content Repurposing In Claude Cowork
Creator
Wilbert Kramer
Published on
aidecode.substack.com (Substack)
Original date
Mar 12, 2026
License
Not openly licensed; the original prompt is shown word for word with credit and a link to the original
Originally for
Claude Cowork
AI dry run
Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. PASS: Gates held on desk check; no live sends or purchases.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails.
Verified
Oct 24, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.

Are you the creator? Claim or correct this page.

More in Marketing

All 97 Marketing playbooks →

Related playbooks

MarketingWorkflow✓ AI dry run

Build a weekly LinkedIn Ads pacing report with Claude

Have Claude turn each week's LinkedIn Ads results into one written report: progress against pipeline targets, the main platform metrics, the audiences, campaigns and ads that are falling behind, ads whose click-through to the landing page is slipping, and a short list of last week's finished work and what is planned next.

Credit: Patrick Cumming

An afternoon+
MarketingPrompt✓ AI dry run

Build a terminal command that edits your tech drafts

Run one command on any Markdown draft and get a structured edit covering typos, clarity, structure, code accuracy and publishing readiness.

Credit: Connie Leung

Under an hour