Challenge

I publish content across X, LinkedIn, Instagram, TikTok, and YouTube for Second Ladder, a career network with 48K+ newsletter subscribers. The strategy is built around the X Thread Waterfall: write one thread, then derive every other platform's content from it. LinkedIn post, IG carousel, standalone tweets, stories, video scripts, all from one pillar piece.

The problem was execution. I'd go weeks (64 days at the worst point) without posting because the research, writing, formatting, and cross-platform syncing took too long to do manually. Every piece required scanning current events, checking what competitors covered, writing in a specific voice, formatting for each platform, saving to Notion, creating an Airtable tracking row, and updating a local content library. Too many steps, too many places for the chain to break.

What I Built

A Content Flywheel: an automated scheduled task that handles the entire pipeline from research to draft-ready content packages. It runs on a schedule, does all the upfront work, then pauses for human approval at three checkpoints before moving forward. The output is a complete multi-platform content package ready to post, with everything already synced to Notion and Airtable.

How It Works

The flywheel runs in three stages with approval gates between each one.

Stage 1: Research and picks. The system scans current events (web search for breaking stories, funding rounds, IPOs, layoffs, earnings), checks what competing newsletters covered (Morning Brew, The Hustle, Exec Sum), reads the research bank and content performance history, then surfaces 6 topic picks. Each pick comes with a lane assignment (AI, Hiring/Firing, Senior Job Search, etc.), hook options, CTA type, recommended publish date, and two alternate topics in case I want to swap. It also flags lane distribution imbalance and stale research. The picks are saved as a markdown file and presented for approval.

Stage 2: Full content draft. Once I approve a topic, the system writes the complete waterfall package: a 10-tweet X thread, 9 standalone tweets for the queue, a LinkedIn long-form post, an 8-slide IG carousel outline, 5 IG stories, a universal caption for all platforms, 3 editorial image prompts, and a 45-60 second talking head video script with recording notes. Everything is written in my documented speaking voice (connector words, present continuous, soften absolutes, reference back before introducing new info). The draft runs through an automated audit against platform-specific prompt docs and the Waterfall System Generator stored in Notion.

Stage 3: Sync and archive. The finished content is saved to a local script file, created as a Notion sub-page under Video Scripts (with full inline hyperlinks preserved), tracked in Airtable ContentOS (title, status, format, content pillar, upload location, brand, internal doc URL, cover image prompt), and logged in the content-library.md memory file. All four destinations update in a single pass.

In this session, I approved two topics and got two complete packages: a thread on the SpaceX/Anthropic/OpenAI IPO wave (10 tweets + full waterfall) and a short script on companies spending more on AI tokens than employee salaries. Both packages were research-backed with 7+ sourced data points each.

The Stack

Claude Cowork scheduled tasks for the automation backbone. Notion MCP for content archiving under an organized hierarchy (Video Scripts parent, month pages, individual entries). Airtable MCP for ContentOS tracking (status, format, pillar, brand, platform, internal doc URL). Web search for real-time event scanning and competitive newsletter checks. The content strategy framework (X Thread Waterfall, 6 content lanes with target percentages, 90-Day Content Challenge) is documented in Notion and referenced automatically during each run. The whole system is repeatable: any new topic goes through the same pipeline and lands in the same places.

Results

Designed to: Eliminate the research-to-publish bottleneck that caused a 64-day content drought. Produce 2-3 complete multi-platform content packages per week with 20-30 minutes of founder time per day (approving picks, reviewing drafts, recording video).

This session: 2 full content packages drafted, audited, and synced in a single run. Each package included thread + standalone tweets + LinkedIn post + IG carousel + IG stories + video script + image prompts + universal caption. All synced to Notion, Airtable, and local content library with zero manual data entry.

Pending real metrics: Post engagement, follower growth, and newsletter signup attribution will be tracked once the content starts publishing. The 90-Day Content Challenge (May 5 to Aug 2, 2026) is the measurement window.