> ## Documentation Index > Fetch the complete documentation index at: https://docs.rippit.com/llms.txt > Use this file to discover all available pages before exploring further. # enrich_worksheet > **Analyze with AI** Add AI columns to a worksheet that classify themes and root causes, score records or extract evidence. Each row's answer uses its full conversation transcript, found through the worksheet's `ticket_sid` column or the column whose kind is `external_gradable_reference`, plus any `contextColumnIds` values; a worksheet with neither reference is judged on its own columns instead. Results persist as columns that can be filtered and aggregated. Only rows already on the worksheet are processed; rows that load after the job starts are not enriched. All columns requested together are filled in a single LLM call per row. The job runs in the background and this returns its `jobId` at once; get_enrich_status with the worksheetId and jobId reports progress and completion. `columns` pairs each new column's display name with its `id`, and queries take the `id`. ## OpenAPI ````yaml /api-reference/tools-openapi.json post /v1/enrich_worksheet openapi: 3.1.0 info: title: Rippit API version: 1.0.0 description: Every operation is also an MCP tool with the same name and parameters. servers: - url: https://api.rippit.com security: - bearer: [] paths: /v1/enrich_worksheet: post: summary: enrich_worksheet description: >- **Analyze with AI** Add AI columns to a worksheet that classify themes and root causes, score records or extract evidence. Each row's answer uses its full conversation transcript, found through the worksheet's `ticket_sid` column or the column whose kind is `external_gradable_reference`, plus any `contextColumnIds` values; a worksheet with neither reference is judged on its own columns instead. Results persist as columns that can be filtered and aggregated. Only rows already on the worksheet are processed; rows that load after the job starts are not enriched. All columns requested together are filled in a single LLM call per row. The job runs in the background and this returns its `jobId` at once; get_enrich_status with the worksheetId and jobId reports progress and completion. `columns` pairs each new column's display name with its `id`, and queries take the `id`. operationId: enrich_worksheet requestBody: required: true content: application/json: schema: type: object properties: worksheetId: type: string minLength: 1 pattern: ^[A-Za-z0-9_-]+$ description: Worksheet to add the columns to. enrichments: minItems: 1 maxItems: 25 type: array items: type: object properties: name: type: string minLength: 1 maxLength: 120 description: >- New column display name, unique within the call ignoring case and treating spaces as underscores. instructions: type: string minLength: 1 maxLength: 10000 description: >- Per-row task for the model. With `overrides`, the default for rows no override matches. outputType: oneOf: - type: object properties: type: type: string enum: - string - array(string) - integer - array(integer) - float - boolean required: - type additionalProperties: false - type: object properties: type: type: string const: enum values: anyOf: - minItems: 2 type: array items: type: string - minItems: 2 type: array items: type: integer minimum: -9007199254740991 maximum: 9007199254740991 description: >- Allowed labels; rows the model cannot judge are forced into one unless a catch-all such as "unknown" is listed. multi: type: boolean required: - type - values additionalProperties: false description: >- Result type; `enum` values are stored as text labels. With `overrides`, the default-case type. overrideCaseColumnId: description: >- Existing pipeline-born column whose row value selects the override; required with `overrides`. Every column in one call that uses overrides must branch on the same column. type: string overrides: description: >- Per-value instructions, one entry per overrideCaseColumnId value; rows matching no entry use the column-level instructions and outputType. minItems: 1 maxItems: 60 type: array items: type: object properties: case: type: string minLength: 1 description: >- The overrideCaseColumnId row value this entry applies to: an enum column's label text, or a bucketed column's `opt_*` id. instructions: type: string minLength: 1 maxLength: 10000 description: >- Instructions for rows whose overrideCaseColumnId matches `case`. outputType: description: >- This branch's output type; omitted reuses the column's. It must keep the default's underlying data type (string enums and plain strings mix). oneOf: - type: object properties: type: type: string enum: - string - array(string) - integer - array(integer) - float - boolean required: - type additionalProperties: false - type: object properties: type: type: string const: enum values: anyOf: - minItems: 2 type: array items: type: string - minItems: 2 type: array items: type: integer minimum: -9007199254740991 maximum: 9007199254740991 description: >- Allowed labels; rows the model cannot judge are forced into one unless a catch-all such as "unknown" is listed. multi: type: boolean required: - type - values additionalProperties: false required: - case - instructions additionalProperties: false required: - name - instructions - outputType additionalProperties: false description: Columns to create and fill. contextColumnIds: description: >- Existing column ids whose per-row values the model reads alongside the conversation. type: array items: type: string where: description: Enrich only rows matching these filters. maxItems: 20 type: array items: $ref: '#/components/schemas/enrich_worksheet____schema0' sampleRows: description: >- Enrich only a uniform random sample of this many rows, drawn after `where` and fresh on each run. type: integer minimum: 1 maximum: 10000 model: description: >- Model id; omitted uses the server default. Choose another supported model when the task calls for it. Ids are tenant-specific, and unknown ones are rejected. type: string minLength: 1 createScaffoldIfMissing: description: >- True turns a worksheet with no pipeline columns into a reusable pipeline on this enrichment; unset runs a plain one-off job. type: boolean required: - worksheetId - enrichments additionalProperties: false responses: '200': description: Tool result. components: schemas: enrich_worksheet____schema0: anyOf: - type: object properties: columnId: type: string minLength: 1 description: Column id, not its label. operator: type: string enum: - '=' - '!=' - '>' - < - '>=' - <= - IN - NOT IN - LIKE - CONTAINS_ALL - ANY - ALL - NONE - NOT_EMPTY - EMPTY - IS - IS NOT description: >- IN/NOT IN match exact values or array elements; CONTAINS_ALL needs all. ANY/ALL/NONE: case-insensitive substring matches, never for id columns (they over-match). value: description: >- Array for multi-value operators (ANY/ALL/NONE: at most 250 strings). singleSelect and bucketed columns store option ids (opt_*): filter by id; a label matches no rows. describe_column with `search` resolves a label to its id. anyOf: - type: - string - number - boolean - 'null' - type: array items: type: - string - number - boolean - 'null' required: - columnId - operator additionalProperties: false - type: object properties: columnId: type: string minLength: 1 description: Column id, not its label. operator: type: string const: BETWEEN date_from: type: string description: >- Inclusive. Plain dates cover whole days, and times without an offset are clock times, in the reported timezone or UTC. date_to: type: string description: Inclusive; a plain date covers that whole day. required: - columnId - operator - date_from - date_to additionalProperties: false - type: object properties: columnId: type: string minLength: 1 description: Column id, not its label. period: type: object properties: 'n': type: integer minimum: -9007199254740991 maximum: 9007199254740991 description: Signed units; negative values reach into the past. unit: type: string enum: - hours - days - weeks - months - years required: - 'n' - unit additionalProperties: false description: >- Whole days in the read timezone, including today: { n: -7, unit: "days" } keeps today and the six prior days. Months/years follow the calendar; hours are exact. Reads re-resolve; worksheets keep a snapshot. offset: description: >- Shifts the anchor before period applies; { n: -1, unit: "weeks" } moves it one week into the past. type: object properties: 'n': type: integer minimum: -9007199254740991 maximum: 9007199254740991 description: Signed units; negative values reach into the past. unit: type: string enum: - hours - days - weeks - months - years required: - 'n' - unit additionalProperties: false required: - columnId - period additionalProperties: false - type: object properties: or_filters: minItems: 1 type: array items: minItems: 1 type: array items: $ref: '#/components/schemas/enrich_worksheet____schema0' description: ORs branches of ANDed items. required: - or_filters additionalProperties: false securitySchemes: bearer: type: http scheme: bearer ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.