Module 1 · What agents are · Lesson 1
What is an agent?
From a single model call to an LLM with tools to a loop that picks its own next step, and how to tell which one a real task needs.
By the end of this lesson you can
- Explain the difference between an LLM call, an augmented LLM, a workflow and an agent, using precise definitions
- Trace an agent loop turn by turn and identify the model's decisions and the harness's work in each turn
- Compute how per-step reliability compounds across a multi-step process, and use that to argue for or against autonomy
- Classify the steps of a business workflow as code, single call, workflow, agent or human-only, and justify each choice