Actions7
- Data Operation Actions
- Agent Operation Actions
- Task Actions
Agent Operation → Agent Map
AI-generatedSummary
Run an AI agent to process each input row by performing a specified research or enrichment task, with configurable effort level and optional response schema, optionally merging the AI output back with the original row data.
Inputs
- Session Name — Optional name for the Everyrow session visible in the dashboard, defaulting to 'n8n-everyrow-session'.
- Task Description (required) — A text description of the task you want the AI agent to perform on each row, such as researching company details.
- Effort Level — Select the research depth the AI should apply per row; options are Low (quick research), Medium (moderate), or High (thorough). Defaults to Low.
- Response Schema (JSON) — Optional JSON Schema defining the structure for the agent's response. Leave empty to use default output formatting.
- Join With Input — Boolean to determine whether each AI agent output should be merged into the original input row data (default true).
Output shape
a single object containing the submitted task ID, session ID, status, and operation type
The operation returns immediately with task tracking information; the AI processing runs asynchronously and results can be retrieved via subsequent 'Get Result' operations using the task ID.
Examples
Example 1: You have a table of companies and want to enrich each row with AI-researched data about the latest funding rounds. You specify a relevant task description, choose an effort level, optionally define a JSON schema for the response structure, and run the operation. The node submits the task and returns an ID you can use later to fetch results.
The node inputs are the array of JSON rows to process, the 'Task Description' specifying the research to do per row, an effort level (e.g. Medium), optional JSON schema for response consistency, and 'Join With Input' set to true to combine AI results into original rows.