Running agents using items
When running agents, you can provide two types of optional inputs: items and context. Understanding the difference helps you control both what the agent works on and how it generates results.
Items are the inputs the agent acts on. The agent processes each item and generates outputs for it. Use items to define what the agent should process individually.
Context is supporting information that guides how the agent generates results, without being directly processed. The agent generates responses grounded in this context. Use context to guide tone, style, or constraints.
Use items when you want the agent to perform actions on specific content.
To add items, click Add items, then select one or more items from the drop-down, and click Save.
Each selected item is processed individually, and the agent generates corresponding outputs.
Types of items
You can select different types of items depending on your workflow:
-
CMS - content items from a SitecoreAI site or collection.
Example: Translate multiple pages from a SitecoreAI site into different languages.
-
Artifacts - artifacts previously generated by agents.
Example: Use the content generated by one agent as input for another. See Run additional agents for details.
-
Briefs - briefs created in the Strategy app.
Example: Use a smart brief to generate campaign content.
-
Other - full or partial data imported from CSV file.
Example: Use selected accounts from a CSV for account-based marketing (ABM).
How items work with agents
When you run an agent with items:
-
The agent processes each item individually
-
Parameters (such as language or format) are applied to each item
-
Outputs are generated per item and added to the space
This makes it possible to scale tasks across multiple inputs in a single run.