AI Generator

This name identifies the AI Generator configuration and is used to label the generated document and related logs.

Best Practices:

• Use descriptive names that reflect the purpose or output of the generator (e.g., "Invoice Summary Generator" or "Contract Key Points").
• The name will appear in user interfaces and reports, making it easier to track and manage multiple generators.

Select the LLM that will be used to generate the output document. The available models are determined by your LLM Connector configuration and may include options such as GPT-3.5, GPT-4, or other OpenAI-compatible models.

Configuration Guidance:

• Choose a model that balances cost, speed, and quality for your use case.
• More advanced models (e.g., GPT-4) may provide higher quality or more nuanced outputs, but may also incur higher costs or longer response times.

Tip: If unsure, start with the default model and adjust as needed based on output quality.

Provide clear, detailed instructions that specify the type of document to create, the content to include, and any formatting or style requirements.

Guidance:

• State the desired output format (e.g., summary, report, list, HTML, JSON).
• Indicate which information from the input documents should be included or emphasized.
• Optionally, describe the intended audience or use case for the generated document.

Example Instructions:

• "Generate an HTML document listing the name, address, and phone number for every person or organization referenced in the document quotes."
• "Summarize the key findings from the attached reports in a bulleted list."
• "Create a JSON object containing the invoice numbers and total amounts for all invoices in the input."

The quality and relevance of the generated document depend heavily on the clarity and specificity of these instructions.

Provide a sample output to guide the AI model toward your preferred format, structure, or style.

Usage:

• Use an example to demonstrate the expected layout, level of detail, or tone.
• Examples are especially helpful when the instructions are complex or when a specific output format is required.

Example: The AI will use this example as a reference when generating the output.

Can be one of the following types:

Controls how content is selected from the input documents to be provided to the AI model.

Options:

• Choose a quoting method to include only relevant sections, highlights, or quotes from the source documents.
• If no quoting method is specified, the entire input document content will be included.

Best Practices:

• Use quoting to focus the AI on the most important or relevant parts of the input, improving output quality and reducing unnecessary context.
• Select or configure a quoting method that matches your document structure and extraction goals.

For more information, see the documentation for Quoting Method.