Building an agent from a template

You can choose from more than 100 AI agents in the templates gallery to kickstart your agent-building process on the Agent Garden home page. These templates provide inspiration through a variety of examples. After you install a template, review and edit the configuration, including guardrails, to ensure the agent meets your business needs. When you are ready, package and deploy the agent to your chosen environment. Refer to Building an agent for more details on building and configuring agent profiles, tasks, and guardrails.

The gallery also includes:

• Marketplace agent templates - Install configurable agents built by Technology Partners and Boomi Community members. Refer to Installing Marketplace agents to learn more.
• Recommended agent templates - Install configurable agents recommended to you based on your account metadata. These templates highlight integrations in your account where adding the Agent step can add value. Refer to Installing recommended agents to learn more.

Templates labeled with the Deploy button are ready to use and do not require any user credentials or tool configuration. Templates with the Configure button require you to provide credentials or need additional setup before you can package and deploy them.

Context window limitations

The user prompt, agent goal, tasks, and instructions combined cannot exceed 200K input tokens. Refer to Agentstudio token limits for details.

To set up and deploy templates that require configuration:

Step 1: Import the template

1. Navigate to the Home page in the Agent Garden.
2. Select Explore Templates Gallery.
3. Optional: Search for a template using a keyword.
4. Click Configure.

Step 2: Configure tools

5. Click Edit.
6. In Tasks, an exclamation point icon indicates that tools in Tasks need configuration.
7. Select Manage Tools in a task that contains an Action Required label.
8. Select the tool name to open it. The tool is currently in draft status.
9. In the tool, select Configuration.
10. In the configuration screen, any authentication fields are left empty for you to add to. For example, in an API tool, the API URL indicates which API is being called, and the <your-api-key> indicates that you need to add your key for the API.
11. Select Review when finished.
12. Click Deploy Tool.

Step 3: Test and deploy the agent

14. Navigate to Agents and open the agent draft you are working on.
15. In the Test agent window, select a runtime cloud to choose which regional cloud runs your test session. Start conversing with your agent to test it and troubleshoot any issues with tool configuration. The agent trace will indicate if authorization issues occur. Refer to Testing and troubleshooting an agent for troubleshooting tips for tool configuration.
16. Select Review.
17. Select Create, then Package & Deploy. Refer to Packaging and deploying an agent for details on linking runtime clouds and setting up environments. Choose your target production environment and select Create Package & Deploy. Your package is assigned a version number automatically. Once deployed, your team can interact with the agent in the conversational user interface. Refer to Using AI agents to learn more.

For details on managing packages, deploying to multiple environments, and working with regional runtime clouds, refer Deploying agents.

note

Your agent always remains in a draft state until you package and deploy it. You can continue editing it at any time without affecting active deployments. Changes only go live when you create and deploy a new package.

Step 4 (Optional): Add agent to Agent step

The Agent step allows you to integrate AI agents from the Agent Control Tower that you create with the Agent Designer directly into your integration process workflows.

You can find the Agent step in the Execute section of the Steps Palette on the Integration Build canvas. After adding it to the canvas, choose from the available agent deployments in your Control Tower to configure in the step. Each entry in the list represents a specific deployment and runtime cloud combination, so you may see the same agent listed more than once if it is deployed to multiple environments. Refer to Agent step to learn the basics and Creating a Data Quality Report agent to see how you can use Agent step in an integration.

Embedding agents in your applications

The Boomi Agentstudio Embed Kit allows you to embed agents you create in Agentstudio directly into your own applications. Get started with the GitHub repository and explore the Boomi Embedded documentation.