Agent
An AI that works through a multi-step task on its own, deciding what to do next and using tools along the way.
What it is
An agent is an AI model running in a loop: it reads the goal, picks a tool (search the web, read a page, look up a record, send an email), looks at the result, and decides the next step until it's done.
That makes agents good at open-ended jobs like researching a company, enriching a lead, or triaging a ticket, where the steps aren't known in advance. A coding agent is one: it reads, writes, and tests code until your feature works.
Agents need clear goals, a limited set of tools, and a defined finished result. Vague goals produce wandering agents.
How to ask for it
“Have AI find information about each lead.”
“When a lead is added, run an agent that searches the web for their company, reads the company site, and returns industry, employee count, and a one-sentence summary. Save the result on the lead and show "Researching…" until it finishes.”
"Find information" has no finish line; describing the agent's tools, output, and status gives it a clear job.
You've seen this in
- DCoding agents that build apps from a description
- AAI research assistants that browse the web
- SSupport bots that look up your order and issue a refund