In this part of my AI Masterclass for absolute beginners (part 9), I break down what AI agents are in the simplest way possible. I use a mental model I really like: think of a “digital intern.” Normal generative AI is like an intern who answers your questions—input goes to a model, and you get output. But an AI agent is the intern that doesn’t just reply; it takes action and owns the task end-to-end. I explain how an agent works as a loop: observe, think, act, repeat—until the final goal is achieved. That means the agent can decide which tool to use, which route to take, which model to invoke, and it can validate results and keep going. I also give a simple example: “book me the cheapest flight.” An agent will search options, compare prices, pick the best one, and continue through booking instead of stopping after one response. My key takeaway: an AI model is the smart brain, but an AI agent is the brain plus hands. Agents can call APIs, access databases, send emails, and update apps like Slack/Discord/CRMs to actually get real work done—not just talk about it.