31 days. 31 AI agents. One build shipped every single day.
In May 2026, as a CREAO Ambassador in Indonesia, I gave myself a public challenge: design and publish one AI agent built with CREAO AI every day for 31 consecutive days. I called it 31 Days of Agent — and I finished all 31.
Why I started
AI is shifting from simple prompts into autonomous systems, workflows and collaborative agents. I wanted to understand that transition hands-on. Instead of only consuming AI content, I wanted to build publicly, think through systems daily, and share the process openly.
- Build publicly — no polishing in secret
- Think through agent design patterns daily
- Test ideas rapidly and iterate fast
- Inspire other builders to start creating
- Document a creative lab for future products
Building in public forces ideas to evolve faster.
What I explored
The 31 agents spread across eight areas:
- Strategy — content strategy, planning systems and decision-support agents
- Analytics — scoring, grading and analyzing content and performance
- Content systems — repurposing and multi-format content generation
- Copywriting — hooks, headlines and CTAs
- Social media — trend detection and content angles for X, LinkedIn and beyond
- AI research — style analysis and pattern extraction
- Productivity — weekly reviews and reflection engines
- Experimental — weird, fun and unexpected agents
Some agents were experimental. Some were useful immediately. Some were weird. Some may eventually become real products. But every day shipped something.
The stack
CREAO AI and its Super Agent, MCP connectors, Google Sheets as the data layer, LinkedIn, X and Discord for sharing, and Netlify for the challenge site.
Not every agent needs to become a startup. Some exist to explore possibility.