Built an AI-powered demo submission pipeline for a dubstep record label - automatically screening artist demos for genre fit, tagging them, and notifying the A&R team via Slack.
Deep Medi Musik receives a high volume of demo submissions from artists. Every track had to be manually reviewed with no structured way to screen, categorise, or prioritise. Demos sat in inboxes and shared folders. The A&R team had no pipeline and no way to quickly filter by genre suitability - meaning good demos got buried and time was wasted on poor fits.
<aside>
Make, Airtable, Gemini, SoundCloud, Slack
</aside>
<aside>
System Architecture & Maintenance
</aside>
The automation handles intake, metadata scraping, AI analysis, database updates, and Slack notifications in a single pipeline. The system originally supported SoundCloud, Bandcamp, and raw file uploads - after Bandcamp locked down their API, the client opted to streamline to SoundCloud-only submissions.
A scheduled Make scenario runs every Monday at 8am, pulling all unreviewed demos from the past week and pushing a formatted summary to the A&R team's Slack channel.
Each message includes the artist's details and an embedded SoundCloud preview - so the team can listen directly in Slack without opening Airtable.
Built with Airtable Interfaces - a clean, tabbed view for the team to review and manage submissions without touching the backend.
