> For the complete documentation index, see [llms.txt](https://docs.rillet.com/EjmEP4KZ9BVp9j2ho4Mh/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.rillet.com/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/aura-agents.md).
# Aura Agents
An Agent is a multi-step plan that Aura builds based on a request you describe. Instead of completing tasks one by one in Aura Assistant, you can review the full plan, make adjustments, and run it as a structured process. Agents are useful for complex tasks, recurring processes, or anything you want to standardize and automate.
### Understand When to Use Agents
Use Agents when you need more structure and control over your tasks:
| |
| ------------------------------------------------- |
| A task requires multiple steps |
| You want to review the full plan before execution |
| You need to repeat the same process regularly |
| You want to schedule work to run automatically |
For simple questions or quick analysis, use Aura Assistant instead.
### Create an Agent
Follow these steps to create an Agent from scratch or use a suggested one:
1. Click **Aura** in the left navigation menu.
2. Click **Agents** at the top of the panel.
3. Type what you want the Agent to do directly into the **Create an Agent** box, then press enter or click the send icon.
4. Or, click **Suggested** to browse prebuilt Agent ideas instead, then select one to use as a starting point.
{% hint style="info" %} **Note:**
You can also reopen a previously saved Agent from **Saved**, or a recent session from **Recents**.
{% endhint %}
Aura generates an Agent plan that includes a list of tasks, their run order, and dependencies between them, and can use any files you include in your request as additional input.
### Review and Run the Plan
Aura opens the plan on a review screen, listing each step as an **Agent Task** card in run order. Click the arrow on any card for its full instructions, or type a change request below the plan to refine it instead of clicking Execute. The buttons below decide what happens next:
| Button | What It Does |
| ----------------------------------------- | ------------------------------------ |
| [**Execute**](#execute-or-dismiss-a-plan) | Runs the plan immediately |
| [**Schedule**](#schedule-a-plan) | Runs the plan at a later time |
| [**Save**](#save-and-reuse-agents) | Saves the plan to reuse later |
| [**Add Tasks**](#add-tasks-to-a-plan) | Adds a new step to the plan |
| [**Dismiss**](#execute-or-dismiss-a-plan) | Discards the plan without running it |
A status line below the buttons shows the plan's current state, such as *Aura AI is requesting your approval for the workflow plan* or *Aura AI is executing your workflow*.
### Execute or Dismiss a Plan
Complete these steps to execute or dismiss a plan:
1. Click **Execute** to run the plan right away and watch Aura show progress for each step as it executes.
2. Or, click **Dismiss** instead to discard the plan without running, saving, or scheduling it.
Executing puts the plan into motion, while dismissing clears it without making any changes.
### Schedule a Plan
Perform these steps to schedule a plan:
1. Click **Schedule** on the plan review screen.
2. Select one of the following options:
* **Tonight at 11:00 PM**
* **Tomorrow at 9:00 AM**
* **Custom…** to pick your own date and time
All schedules run on your subsidiary's timezone, always reflecting the default subsidiary you're working in, and you can pause, edit, or cancel a schedule at any time.
### Save and Reuse Agents
Take these steps to save an Agent:
1. Click **Save** on the plan review screen.
2. Enter a **Name** and a **Description**.
3. Choose whether to **Share with others in your org**.
4. Click **Save** to confirm.
Saved Agents appear under **Saved** and can be reused at any time.
### Add Tasks to a Plan
Use these steps to add a task to the plan:
1. Click **Add Tasks** on the plan review screen.
2. Enter a **Task Name**.
3. Enter **Instructions** describing what the task should do.
4. **(Optional):** Set **Input Sources**, the tasks this one depends on, or **Feeds Into**, the downstream tasks that depend on this one.
5. Click **Add** to add the task to the plan.
The new task appears in the plan alongside the others, ready to run with the rest.
### Frequently Asked Questions (FAQs)
Do Agents need my approval before taking action?
Yes, in the same cases Aura Assistant does. Whenever a step in the plan creates, updates, or sends data, the Agent pauses at that step and shows an approval card so you can approve, edit, or reject the action before it continues from that same step. Read-only steps, like analysis or reporting, run automatically without needing your approval.
Do scheduled Agents skip approvals?
No. Scheduled Agents do not bypass approvals. If a scheduled run includes an action that modifies data or sends communication, it pauses and is marked **Needs attention** until you review it, then resumes and continues execution once you approve, edit, or reject the action, keeping every change under your control even when the Agent runs automatically.
Can I see what an Agent did in past runs?
Yes. Every Agent run is recorded, so you can go back and review previous executions, see the results and outputs they produced, and check which actions were approved or rejected. This makes it easy to track performance over time and understand how your Agents are actually being used.
When should I turn a task into an Agent?
If you find yourself repeating the same task regularly, that's a good sign to turn it into an Agent and schedule it. It cuts down on manual work and helps keep the result consistent every time.
### **See Also**
To learn more about Agents in **Rillet**, review these articles:
* [**Getting Started with Aura Assistant**](/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/getting-started-with-aura-assistant.md)
* [**Upload Files to Aura**](/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/connect-external-apps-with-aura-mcp-connector/upload-files-to-aura.md)
* [**Aura Actions and Approvals**](/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/aura-actions-and-approvals.md)
* [**Audit Trail**](/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/audit-trail.md)
* [**Assistant vs Agent**](/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/assistant-vs-agent.md)
* [**Aura AI Prompt Examples**](/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/aura-ai-prompt-examples.md)
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.
Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:
```
GET https://docs.rillet.com/EjmEP4KZ9BVp9j2ho4Mh/aura-ai/aura-ai/aura-agents.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.
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.
