AI Agents in Action.
A chatbot answers. An agent gets work done — it takes a goal, makes a plan, uses tools, checks the result, and loops until it's finished. This is the fastest-growing thing in AI. It's also where most people faceplant. So we're doing it the honest way: the durable fundamentals + the real 2026 tools, so you build agents that actually work — and know exactly when not to.
By the end, you'll be able to
- Tell an agent apart from a chatbot and a workflow — and pick the simplest thing that works
- Explain agent anatomy (tools, memory, the reason→act→observe loop)
- Build a single agent no-code, with a real system prompt + knowledge base
- Navigate the 2026 platform landscape and choose the right tool
- Design a multi-agent system — and know when it's overkill
- Evaluate quality, add guardrails & governance, and control cost
- Ship your capstone agent + earn your certificate
For makers & operators ready to go past prompting. The sequel to AI on Autopilot. Anti-hype, receipts included.
You're an Agent Operator.
First, the capstone. Tick off each piece as you ship your agent — start with Capstone A (a content-repurposing agent); ambitious baddies also build Capstone B (a multi-agent research pipeline).
Then the knowledge final — 10/12 (80%) to earn your certificate. Retake as needed.