QuoteFormat

Use this format to inject machine-readable data that the AI can easily parse and interpret. The output preserves the full structure of the Data Model as a JSON object, making it suitable for advanced analysis, validation, or integration scenarios where structured data is required.

Example

{
  "Invoice Number": "INV-12345",
  "Total Amount":  "1,234.56"
}

Best Practices

• Use JSON when the AI workflow expects structured input or when downstream processing requires precise data extraction.
• Review the generated JSON artifact to ensure it matches the expected schema and content.

Use this format for a more human-readable presentation of the data, especially when quoting paragraph-style or narrative content that benefits from being laid out like a document. The output uses Markdown headings and formatting to organize fields and values, improving clarity for natural language content.

Example

In this example, vendor information and coding instructions are injectd in Markdown format, to help the LLM assign GL accounts to each invoice line item. These fields were populated through a previous extraction step.

### Vendor Name

Office Depot

### Coding Instructions

Code products as Office Supplies.(Account No: 56789) and services as Consulting.(Account No: 98765) 

Best Practices

• Use Markdown when the AI benefits from a natural language presentation of the data.
• Suitable for summarization, review, or scenarios where data structure is less critical than presentation.