AI Agents Intro
Before you build with Pluto, it helps to be clear on what an AI agent actually is—and how it differs from the chatbots you've used before.
An AI agent is an AI that—in addition to answering questions—it takes actions. It works autonomously to accomplish a task you assign it, planning its approach, making decisions, and using tools to carry those decisions out.
More Than a Chatbot
A traditional chatbot is reactive. You ask it something, it replies with text, and the exchange ends there. It can tell you how to do something, but it can't actually do it for you.
An agent closes that gap. Given a goal, it can look things up, call external services, transform data, and chain several steps together—all on its own—until the task is done. The conversation becomes a means to an outcome, not the outcome itself.
How an Agent Works
Every time you hand an agent a task, it moves through three things:
- Plans — It breaks your request into the steps required to fulfill it, and works out the order they need to happen in.
- Decides — At each step it chooses what to do next based on the goal, the context of the conversation, and the results it's seen so far.
- Uses tools — When answering isn't enough, it reaches for the tools it's been given—APIs, data sources, and integrations—to actually execute the action.
Because it can reason and act in a loop, an agent handles open-ended, multi-step work that a simple request-and-response bot never could.
Where Pluto Fits In
Pluto gives you a production-ready AI agent that already knows how to plan, decide, and use tools—so you don't have to build any of that from scratch. This base agent is designed to be extended. It is the foundation on which you build your own agent.
To build on the base agent; you name it, define its purpose, provide instructions and configure its tools. All this is done in plain language, no coding required.
Ready to see it in action? Head to the Quickstart Guide and you'll have your own agent running in under 15 minutes.