Structure & Content

1. Structure & Content

• Data sources & creation → How a Sheet is created (scratch, CSV, AI).

• Columns & fields → Supported column types and field formats.

• Rows & records → Adding, editing, and bulk uploading rows.

• AI skill columns → Special column type for automated AI operations.

---

Data Sources & Creation

AI Sheets can be created in multiple ways depending on the data source.

Method	Details
Blank Sheet	Start with an empty grid. Columns and rows can be added manually.
CSV upload	Upload a CSV file to create a Sheet with thousands of rows at once. Schema detection automatically maps headers to columns. Unsupported file types (e.g., XLSX, PDF) cannot be uploaded.
AI generation	Create a new Sheet by prompting AI. The system generates column headers and starter rows based on the request.
Default location	New Sheets are stored at the workspace level. If created inside a folder or project, they inherit that parent location.
Schema detection	On CSV import, AI Sheets detect field types (e.g., text, number, date). Invalid or mismatched fields trigger error states, which can be resolved manually.

Columns & Fields

Columns define the structure of an AI Sheet. Each column has a specific type that controls the data it accepts and how it behaves.

Column Type	Details
Text	Free-form text entry. Accepts plain text, links, or short notes.
Number	Stores numeric values. Supports integers and decimals.
Date	Stores calendar dates. Can be used for due dates or scheduling inputs.
Dropdown	Fixed list of selectable values. Dropdown options must be defined at column creation.
AI Skill	Special column type that runs AI-powered operations across rows. Cannot be converted from an existing column — it must be added as new. Outputs depend on the skill definition.

Schema rules

• Columns created by CSV upload are automatically typed (text, number, date) when possible.

• Column types can be changed after creation, except for AI Skill columns, which must always be newly added.

Rows & Records

Rows represent individual records in an AI Sheet. Each row follows the defined schema of the columns.

Feature	Details
Row creation	Rows can be added manually one at a time, through CSV upload, or generated by AI.
Bulk upload	CSV import supports creating thousands of rows at once. Each row maps to the column schema detected during upload.
Editing rows	Any row can be edited directly in the grid. Field validation applies (e.g., numbers must be numeric, dates must be valid).
Row deletion	Deleting a row permanently removes it from the Sheet. Deleted rows cannot be restored unless undone immediately.
Row execution	Rows can be run through AI Skill columns individually or in bulk. Users may choose to run all rows, selected rows, or a single row.
Limits	AI Sheets support thousands of rows per Sheet. Performance may vary with very large datasets.

AI Skill Columns

AI Skill columns allow rows in a Sheet to be processed automatically by Motion’s AI.

Feature	Details
Definition	An AI Skill column is a special column type that runs an AI-powered operation across each row. Skills are defined by prompts or configured actions.
Creation	Must be added as a new column. Existing columns cannot be converted into AI Skill columns.
Inputs	Each row provides its column data as input context for the skill. AI uses the full row when generating output.
Outputs	The AI writes its result back into the AI Skill column for each processed row. Outputs are text by default.
Execution modes	Single row: Run AI only for the selected row. All rows: Run AI across every row in the Sheet.
Progress tracking	While running, a status indicator displays progress row by row. Completed rows update as outputs return.
Error handling	Rows that fail to run display an error icon. Double-clicking reveals the error message. Errors can be retried individually.
Context inheritance	AI Skill columns run using the workspace context. They cannot run using private or external user data unless an integration is configured.