# How Agents Work

**The Agent Process**

When an agent is triggered, it follows a structured reasoning cycle to complete its task:

**1. Task Input**\
The agent receives clear instructions in plain language—for example, *“Find all overdue invoices and email the client.”*

**2. Decision Making with LLM**\
The large language model determines the next step, which could be:

* Responding directly to the task,
* Asking for more information, or
* Activating a connected tool such as a data source, workflow, email, or even another agent.

**3. Tool Execution**\
The chosen tool performs its action, and the results are sent back to the LLM for further processing.

**4. Agentic Loop**\
This cycle of decisions and tool executions repeats until:

* The agent reaches its maximum allowed steps, or
* The LLM concludes that the task is complete.

Each complete cycle—from input to reasoning to output—is called a **run**.

{% hint style="info" %}
⚡ **Key Point:** The LLM’s role is limited to either calling tools repeatedly or providing the final response.
{% endhint %}


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.jetadmin.io/agents-1/how-agents-work.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.