What is an AI agent, and when is one useful?
An AI agent uses a model to choose actions and tools while pursuing a task. It is useful when the path cannot be fully prescribed in advance. A fixed workflow is often easier to operate when the steps, permissions, and success conditions are already known.
What does the source say?
Anthropic distinguishes predefined workflows from agents that direct their own process and tool use. The distinction is architectural; it does not establish that an agent is more valuable for every task.
How can you apply this?
Describe one task twice: first as a fixed sequence, then as a goal with choices. Circle the points that genuinely require judgment. If those choices are narrow, try a small workflow before building a general agent. Set a maximum action budget and an explicit handoff for uncertain or unsupported requests. Use the simplest version that meets the task, and keep a written reason for every permission you add.
What should you watch out for?
Do not equate autonomy with usefulness. Unnecessary tool calls increase cost and failure opportunities without necessarily improving the customer’s outcome.