What is an AI Agent?
A support operation that runs without you having to run it.
Written by Outlearn Documentation
Updated at April 24th, 2026
Table of Contents
Most support teams spend the majority of their time on questions they have already answered before. The same how-do-I questions, the same troubleshooting steps, the same links sent a hundred times. An AI agent changes that equation. It handles the questions that do not need a human so that the humans on your team can focus on the ones that do.
In Outlearn, the agent is the core of everything. It is what your customers talk to. It reads their questions, searches your content, takes action when needed, and knows when to bring a real person in. Everything else you configure in Outlearn - your sources, your actions, your handoffs, your deploy settings - exists to make the agent better at its job.
In this article, you'll learn:
- What an AI agent is and what it does
- What makes up an agent
- How agents relate to the rest of Outlearn
What an Agent Does
When a user starts a conversation, your agent is what responds. It searches through everything you've connected to it, finds the best answer, and replies - instantly, automatically, and around the clock.
But an agent is more than a question-answering machine. Depending on how you configure it, it can also:
- Perform tasks during a conversation (like checking availability or sending an email)
- Recognize when a conversation needs a real person and hand it off seamlessly
- Adjust its tone, personality, and behavior based on the instructions you give it
What Makes Up an Agent
Every agent in Outlearn is made up of the same five components:
| Component | What it does |
|---|---|
| Instructions | Defines your agent's personality, tone, and how it handles different situations |
| Sources | The content your agent learns from and uses to answer questions |
| Actions | Tasks your agent can perform during a conversation |
| Handoffs | Rules for when and how to transfer a conversation to a human |
| Deploy | The channels where your agent is active (website, email, Slack, etc.) |
New to these concepts? Read How Outlearn Works for an overview of all five.
Agents Are Independent
Each agent you create in Outlearn is completely self-contained. It has its own sources, its own actions, its own settings - nothing is shared between agents unless you set it up that way.
This makes it easy to run multiple agents for different purposes side by side. For example: a public-facing customer support agent on your website and a separate internal agent for your team.
To learn how to create and manage multiple agents, see Managing Multiple Agents
Best Practices
- Give your agent a name and instructions that match the context it's deployed in - a customer support agent and an internal HR agent should feel and sound different.
- Think of your agent as a new team member: the better you brief it, the better it performs.
- Start simple. One agent with good sources will outperform a complex setup with poor content every time.