Structured content extractor
The Structured Content Extractor is an agent that analyzes uploaded content and extracts key elements into a standardized, machine-readable format.
It identifies specific structured information such as audience characteristics, key themes, or other defined data points based on the extraction configuration and schema.
After you provide a content source, the agent:
-
Analyzes the content based on the configured extraction type.
-
Extracts relevant structured data.
-
Generates the output as a CSV or JSON file aligned with your defined schema.
When to use
Use the Structured Content Extractor for the following business scenarios:
-
Audience extraction from campaign briefs.
-
Structured data creation for personalization systems.
-
Automated metadata generation for content repositories.
Edit the configuration
Before running this agent, review and update its configuration to control how it generates content.
The configuration determines what the agent produces and how it behaves, so you can align it with your brand, audience, and use case.
Make sure you have a Builder License assigned to your user account. This license is required to edit agents.
-
In the navigation menu, click Agentic. Then, in the left sidebar, select Agents.
-
On the Agents page, find Structured Content Extractor.
-
Click the actions menu
> Edit. -
Review and configure the tabs:
-
Overview - view the agent's basic information, how it works, and inputs required to run it. This tab is read-only.
-
Parameters - define guidance that shapes how content is generated. For example, you can specify target audiences, channels, languages, and formats.
-
Schemas - define structured JSON output formats. For example, you can add schemas for blog posts, email content, or social media posts.
-
Audience Extraction Schema - a predefined brief schema is provided. You can edit or add properties (for example, audience needs, intended audience, age range to match your needs.
-
-
HTML Templates - define how generated content is formatted and displayed in previews. Templates typically correspond to the defined schemas.
-
Workflow - view the agent’s execution workflow, including all processing steps. This tab is read-only.
-
-
Click Save Changes. Return to the Agents page and run the agent.
Run the agent
Use the Structured Content Extractor to analyze provided unstructured content and extract relevant information.
To run the agent:
-
In the navigation menu, click Agentic. Then, in the left sidebar, select Agents.
-
On the Agents page, find Structured Content Extractor.
-
Enter a Space Title for the run.
In the Agents section, the selected agent is added by default.
-
Provide the required inputs:
-
Content to Analyze - enter or paste the content that you want to analyze.
-
-
Optionally, provide the following:
-
Context - provide additional background that guides how the AI generates responses. Click Add context to include:
-
Artifacts - use any artifacts previously generated in the current space.
-
Brand kits - select your organization's brand kit.
-
-
Supporting files - click Browse to upload supported files related to your request.
-
-
Click Create.
-
The new space created for the run opens. Click Run to start the agent.
TipProcessing might take some time. You can leave the page and return to the space at any time to track progress.
When processing is complete, the results appear in the space.
ImportantAlways review AI-generated artifacts carefully and validate key information before using them.
To learn more about what’s included in a space, see View a space.
-
The generated artifact appears with the extracted content, including the following sections, based on the agent's schema:
TipTo customize these sections, edit the agent’s configuration.
-
Audience needs - the audience’s goals, expectations, and the problems the content should address.
-
Intended audience - identifies who the content is for.
-
Audience characteristics - key attributes such as location, age range, interests, and professional background.
-