Case Study

The Manager's Playbook

Content production pipeline for a music-industry podcast — Frame.io, transcription, Cloudflare, Descript, zero manual handoffs.

Client The Manager's Playbook
Industry Media & Podcast Production
Services Workflow Architecture, AI Orchestration, Automation

The Challenge

Podcast production ran on manual handoffs. Raw footage got uploaded to Frame.io, then someone had to manually push it into Descript for transcription. Receipts sat buried in Slack threads instead of the right Drive folder. Signed release forms had to be tracked down and re-sent by hand. When a transcript got re-edited, timestamps drifted, and every flagged moment had to be re-located in the new draft one at a time. Social content — quote posts pulled from episodes — was built image by image in Canva.

The Approach

Starting with a kickoff call to scope the full engagement, we built a connected automation layer on top of the tools the team already used — Frame.io, Descript, Slack, Gmail, Canva, Google Drive — using Claude and Make.com. The core of it is a Frame.io → Descript pipeline (Frame.io webhook → Cloudflare Worker → R2 staging → Descript API) that triggers transcription automatically the moment new footage lands, confirmed working up to 2GB per file. Alongside it: a receipt organizer that reads Slack uploads and routes them to the correct Drive folder, a release-form automation that sends signed forms straight to the right inbox, a timestamp re-mapping tool that finds where flagged moments land after a transcript gets re-edited, and a Canva automation that turns episode quotes into branded Instagram posts without manual design work.

The Result

Five categories of previously manual work — transcription, receipts, release forms, timestamp tracking, and social content — now run without anyone touching them by hand. What started as a single kickoff call scoping the engagement turned into a standing automation suite covering the team's day-to-day production and admin workflow, refined over ongoing sessions as real edge cases came up, like large-file transcription and transcript drift after edits.

← Back to work