Pick & scope
Choose one problem from your real work. Write the one-line user query in the actual user's wording. List the human-review point and what you'd never automate. Share the scope publicly in 4 lines.
Pick one real problem. Build the smallest workflow that solves it. Share progress weekly. Finish with a published workflow that future builders can fork — and a spot in the newsletter.
You can join any time. The 4-week clock starts the day you confirm by WhatsApp or email.
Choose one problem from your real work. Write the one-line user query in the actual user's wording. List the human-review point and what you'd never automate. Share the scope publicly in 4 lines.
Draft the steps. Strip PII from every example. Generate a sample output on real (sanitized) input. Share the draft, including the parts that didn't work.
Get 1-2 peer reviews (in the WhatsApp challenge group or via email). Update steps based on feedback. Highlight what changed and why.
Submit your workflow via /submit-workflow/. After human review it goes live on /workflows/<your-slug>/. Write a 6-line reflection: what worked, what failed, what's next.
Light-touch, but the safety rules are non-negotiable.
Real prompts from the challenge inbox. Pick one or bring your own.
Ship a 30-minute weekly metrics update workflow. 6 numbers + 1 failure line. See the worked example →
Ship a one-source-to-four-channels repurposing workflow that keeps your voice. See the worked example →
Ship a paper-triage workflow with claim-level citation verification. See the worked example →
Ship a non-clinical call routing workflow with an explicit "never automate" list. See the worked example →
Ship a lead qualification workflow that hands every "now" tier to a human within 24 hours. See the worked example →
Ship a community onboarding workflow that escalates to a human within 1 message. See the worked example →
"I'd like to join the 4-week challenge. Audience: ___. Workflow I'll build: ___." That's it.