AnyoneLearnAI
Start learning

Chapter C · 8 pages · ~117 min

What are agents?

Understand agent loops, tools, state, recovery, controls, and evaluation well enough to decide when an agent is the right architecture.

Start the first lesson →Chapter index

What you will be able to do

  • Distinguish an agent from a one-shot model call and a fixed workflow
  • Specify an agent loop using goals, observable state, tools, and success criteria
  • Design bounded planning, memory, recovery, security, and human-review controls
  • Trace a research agent through concrete actions, observations, and failure branches
  • Evaluate reliability, cost, and latency and recognize when not to use an agent
  • Sketch a production architecture that can later connect to multi-agent systems and MCP

Lessons in this topic

  1. 01Agents, workflows, and one-shot calls13m
  2. 02The loop: state, goals, and stopping14m
  3. 03Planning and tool use14m
  4. 04Worked trace: research with evidence16m
  5. 05Memory, context, and recovery14m
  6. 06Limits, security, and human control15m
  7. 07Evaluate agents—or decline them15m
  8. 08Mastery: architect a bounded agent16m