Case Study

The Pattern: Building a Daily AI Culture Briefing

How one non-coder built a fully automated culture intelligence briefing that publishes every morning at 7am -- using RSS feeds, Claude AI, voice cloning, and zero traditional code.

By Mike Litman • Cultural Capital Labs • 2025-2026
50+ RSS feeds
Claude Haiku
ElevenLabs TTS
GitHub Actions
Daily
Updates at 7am UTC
50+
RSS Sources
AI+Voice
Claude + ElevenLabs
Podcast
RSS for Spotify/Apple

The Problem

Culture moves fast. Really fast. Across music, design, fashion, food, technology, and brands, there are hundreds of signals every day -- and no single source captures the full picture. You either spend hours scanning dozens of feeds and feeds, or you miss things. There is no culture equivalent of a morning news briefing.

The existing options are either too niche (focused on one domain), too slow (weekly or monthly roundups), or too algorithmic (serving you more of what you already know). What was missing was a daily synthesis -- something that looks across all of culture and tells you what matters today, before it becomes obvious tomorrow.

The Approach

The Pattern is built on a simple but powerful pipeline. Every morning at 7am UTC, a GitHub Actions workflow triggers a chain of automated steps:

50+ RSS feeds
CultureTerminal data
Claude Haiku synthesis
ElevenLabs audio
Static HTML
Netlify deploy

The source data comes from CultureTerminal, another product in the Cultural Capital Labs portfolio, which aggregates and scores content from over 50 RSS feeds across fashion, brands, design, music, art, and technology.

Claude Haiku (Anthropic's fast, efficient model) takes the raw data and synthesises it into a structured briefing: a Culture Pulse score, a lead story, five signals, a pattern observation, one to watch, and conversation starters. The AI is not generating opinion -- it is finding the pattern across the noise.

ElevenLabs then converts the written briefing into audio using a voice cloned from 30 minutes of training data. The result is a podcast-ready audio briefing that sounds natural, not robotic.

The entire process -- from data fetch to live deployment -- runs unattended. No human intervention required. The briefing publishes at 7am every morning.

How It Was Built

1
RSS Aggregation
Built CultureTerminal as the upstream data source -- an RSS aggregator pulling from 50+ feeds across fashion, design, music, art, food, and technology.
2
AI Synthesis Pipeline
Created a Python pipeline that fetches CultureTerminal data, sends it through Claude Haiku for synthesis, and outputs structured JSON with a Culture Pulse score, lead story, signals, and patterns.
3
Voice Cloning
Added ElevenLabs text-to-speech with a voice cloned from 30 minutes of training data -- producing natural-sounding audio briefings, not robotic output.
4
Automation
GitHub Actions cron job triggers at 7am UTC daily -- fetching data, running synthesis, generating audio, building static HTML, and deploying to Netlify. Zero manual steps.
5
Podcast Distribution
Generated an RSS feed conforming to podcast standards, enabling distribution via Spotify, Apple Podcasts, and any podcast app.
Technical Architecture
DATA SYNTHESIS OUTPUT DISTRIBUTION [50+ RSS Feeds] --> [Claude Haiku] ---> [Static HTML] ---> [Netlify] | | | | CultureTerminal AI synthesis ElevenLabs TTS thepattern.media 5 culture domains Structured JSON Cloned voice RSS podcast feed | GitHub Actions 7am UTC daily

The Result

50+
Sources
Daily
Cadence
7am
Publish Time
0
Manual Steps

The Pattern produces a daily culture briefing covering six sections: Culture Pulse (an overall score), The Lead (the day's top story), 5 Signals (emerging patterns), The Pattern (the connecting thread), One to Watch (a person or brand to follow), and Conversation Starters (talking points for the day).

The audio version has been submitted to Spotify and Apple Podcasts, making it accessible as a daily podcast. An RSS feed powers syndication.

What started as an experiment in AI synthesis became the strongest demonstration of what one person with AI tools can build -- a product that would typically require a content team, an audio producer, a developer, and a product manager.

The Key Insight

AI does not replace taste -- it amplifies it. The curation of sources matters more than the synthesis. What you choose to feed the machine determines what comes out.

The most important decisions in building The Pattern were not technical. They were editorial. Which 50+ sources to include. What structure makes a briefing useful rather than just interesting. How much synthesis is helpful versus how much strips the nuance away.

The technology is the enabler, not the differentiator. Anyone could build this pipeline. The value is in the editorial framework -- knowing what matters, what connects, and what to ignore. That is strategy. That is taste. That is the work.

Lesson Learned
A daily culture intelligence briefing, fully automated, with a cloned voice -- built by one person in a weekend. That is the future of media.
Visit The Pattern Back to All Projects