Definition |
Top Previous Next |
|
The definition tab is where general information about the definition is maintained.
ID is the identifier of the definition. This must be a valid file system name, as the definition is stored at the server in the idp folder under this name.
Description is for informational purposes.
Definition is inactive can be checked to ensure this definition does not process any inbound source documents. This may be used in cases when a job is under development or is used as a template for other definitions.
Auto-transfer can be checked to cause an auto-transfer, which includes post processing, if an inbound document passes all validations defined in the definition. If turned on, full automation is achieved for documents that do not require user review to correct validation errors.
Enable subid overwrite will enable replacement of duplicate subids when the transfer occurs and the document and its subids are written to an archive library. If not checked, sequenced subids are generated.
Instructions is used to provide general instructions in the AI prompt, to indicate some context about what the definition is designed to capture.
AI Model is the default model to process documents with. For example, Claude Haiku.
Alternate Models is a list of alternative models that users will be able to reprocess documents with. Generally, the default model should be the lowest cost model that handles the majority of documents well, and alternate models provide users with higher-level options to use if needed.
Prompt Type is a selection of prompt type (or structure) to use when building a prompt for the AI service. Different prompt structures can be designed for different types of prompts. For example, there could be a type for transaction documents, and a different type for expense documents that don't require line item detail to be captured.
|