An AI system capable of autonomously performing a sequence of tasks toward a defined goal — such as researching a topic, drafting content, and checking it against brand guidelines — with limited or no step-by-step human direction at each individual step, as distinct from a single-prompt AI tool that produces one output per request.
Why it matters
AI agents represent a meaningful step beyond simple prompt-and-response AI tools, capable of handling multi-step marketing workflows with more autonomy — which raises the stakes on human-in-the-loop review (Part Two, Category 9), since more of the process happens without a person checking each individual step along the way.
In a sentence
"We're experimenting with an AI agent that drafts, checks, and formats a first version of routine social posts — but every post still goes through human review before it's scheduled to actually publish."
Related terms
- Often confused with: Prompt / Prompt engineering (Part One and Two, Category 9) — a single prompt produces one discrete output from one instruction; an AI agent can execute a multi-step sequence of actions autonomously toward a broader goal, well beyond a single prompt-response exchange).