Documentation for AI-powered text extraction and transformation functionality
The AI Column Extraction feature allows users to create new columns by processing existing text data using AI language models. Users can extract specific information, transform content, or generate new insights based on source text.
AI-FeatureDemo-480p.mov
- New Column Name: Define the name for the column that will store the AI-generated results
- LLM Provider: Select the configured AI provider with pre-defined parameters - Model, Server URL, Temperature, Max Tokens
- Response Format: Choose the output structure
Response format | Description |
---|---|
`Text | Plain text response |
`JSON Schema | Structured response following a defined schema |
`JSON Object | Free-form JSON response |
- Description: Detailed instructions for the AI model about the desired extraction or transformation
- JSON Schema (Optional): Required when Response Format is set to JSON Schema. Defines the structure and validation rules for the output
Note: There are few manual edits that may be required. Refer to the JSON schema in the example below.
Shows sample processing using the first row of data Displays:
- Input Value: Original text from source column
- Response: AI-generated result based on configuration
Allows validation before processing entire dataset
Use Case | Instruction |
---|---|
Basic Summary | Generate a one-sentence summary of the input text |
Language Translation | Translate the text to German |
Use Case | Instruction | JSON Schema |
---|---|---|
Person information extraction | Extract in JSON format details of the person mentioned in the content | { "name": "individual_schema", "schema": { "type": "object", "properties": { "name": { "type": "string", "description": "The name of the individual." }, "dateofbirth": { "type": "string", "description": "The date of birth of the individual." }, "placeofbirth": { "type": "string", "description": "The place where the individual was born." }, "dateofdeath": { "type": "string", "description": "The date of death of the individual." }, "placeofdeath": { "type": "string", "description": "The place where the individual died." } }, "required": [ "name", "dateofbirth", "placeofbirth", "dateofdeath", "placeofdeath" ], "additionalProperties": false }, "strict": true } |
Use Case | Instruction |
---|---|
Content classification | Classify the content and extract key entities, the response should be in JSON format |
- Summarization
- Translation
- Style conversion
- Format standardization
- Entity recognition
- Key fact extraction
- Timeline creation
- Relationship mapping
- Sentiment analysis
- Theme identification
- Category classification
- Key concept extraction