> For the complete documentation index, see [llms.txt](https://docs.harmony.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.harmony.io/ai-agents/getting-started/working-with-ai-agents.md).
# Working with AI Agents
### Understanding AI Agents
Harmony AI agents are automated workflows that perform tasks such as software compliance checks, employee onboarding, device recovery, app access requests, password resets, and more. Agents are configured per workflow, can run on a schedule or manually, and may require approval before executing certain actions.
#### Agent Types
**Software Management**
Application Access Request
Automates application access requests with approval workflow and identity provider provisioning.
License Expiration Monitoring
Tracks software license expiration dates and creates tickets when licenses are nearing expiration.
Software Compliance Report
Runs periodic software access compliance reviews with manager acknowledgment.
**Access Control**
Device Recovery
Automates device recovery requests with identity verification and secure recovery key delivery.
Password Reset Request
Automates password resets for identity provider users with identity verification.
MFA Device Reset
Automates MFA resets for identity provider users with identity verification.
Group Membership Management
Manages identity provider group assignments with approval workflow.
**Hardware Management**
Device Uptime Monitoring
Identifies devices that have been running for an extended period without a restart.
Low Storage Device Monitoring
Finds devices with critically low free storage and creates tickets for IT remediation.
**Employee Lifecycle**
Employee Onboarding
Automates employee onboarding by creating parent and linked tickets for new hire preparation.
Employee Account Termination
Automates employee offboarding with approval workflow and identity provider account deactivation.
***
### Managing Agents
#### Navigating the Agents List
Open **Agents** from the sidebar to view the list. The Agents page displays all available agents organized by category with the following features:
* **Filters button** - Filter agents by category, status, or approval requirements
* **Search bar** - Search agents by name or description
* **View toggles** - Switch between card view and table view
* **Refresh button** - Reload the agents list
Each agent card shows:
* **Agent name and icon** - Identifies the agent type
* **Category** - Software Management, Access Control, Hardware Management, Employee Lifecycle, or Compliance
* **Status badge** - Enabled (green) or Disabled (gray)
* **Description** - Brief explanation of what the agent does
* **Approval requirement** - "Needs approval" or "No approval needed"
* **Run statistics** - Number of runs and last run time
* **Integration icons** - Shows which integrations the agent uses
#### Viewing Agent Cards vs Table View
Toggle between **Card layout** and **Table layout** to change how agents are displayed:
* **Card view** - Shows larger cards with agent details, descriptions, and visual status indicators
* **Table view** - Displays a compact data table with sortable columns for quick scanning
#### Viewing Agent Details
Click an agent card to open its detail page. The page has tabs: **Overview**, **Runs history**, and **Provisioning** (when applicable). The Overview tab shows the agent's workflow triggers, steps, and configuration panel.
***
### Running and Monitoring Agents
#### Running Agents Manually
On the agent detail page, click **Run** to open the run dialog. Fill in the required inputs (e.g., employee, device, application) and click **Run** to start the workflow. The dialog varies by agent type.
#### Understanding Agent Runs
Each run has a unique Run ID, status, subject, duration, requester, and start time. Runs are triggered by schedules, events, or manual execution.
#### Viewing Agent Run Details
Click a run in the Runs history table to open the run details page. The page shows run metadata, step-by-step execution status, and outputs. Expand steps to see inputs, outputs, and errors.
#### Viewing Agent Execution History
On the agent detail page, open the **Runs history** tab. The table lists past runs with Run ID, Status, Subject, Duration, Requester, and Start Time. Use search and the status filter to find specific runs. Sort by column headers.
#### Understanding Agent Run Status
| Status | Description |
| ------------- | -------------------------------- |
| **Running** | Execution in progress. |
| **Completed** | All steps finished successfully. |
| **Failed** | One or more steps failed. |
| **Aborted** | Run was cancelled by user. |
| **Time out** | Run exceeded the time limit. |
| **Skipped** | Run or step was skipped. |
#### Monitoring Active Agent Runs
Runs with status **Running** show a live duration timer in the Runs history. Open the run details page to see which step is running and monitor progress in real time.
#### Aborting Agent Runs
On the run details page, for runs that are still **Running**, click **Abort** to stop execution. Confirm in the dialog. Aborting stops all remaining steps, actions already completed in external systems may need manual review.
***
### Exporting Agent Data
#### Exporting Agent Run History
To export agent run history to CSV:
1. Open the agent detail page
2. Click the **Runs history** tab
3. Apply any filters or search criteria you want (optional)
4. Click **Export** in the toolbar
5. The CSV file will download with the format `{agent-slug}-runs-YYYY-MM-DD.csv`
---
# 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 dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://docs.harmony.io/ai-agents/getting-started/working-with-ai-agents.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
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.