IDP PROCESSING

Intelligent Document Processing (IDP) utilizes an AI service to perform document data extraction into user-defined data structures used for both archive storage of the documents and external processing.  It is similar in purpose to Image Manager, but performs its actions based on classes of documents rather than formats of documents.  There are two components to IDP processing:

•IDP Definitions, located under Coding Tools on the menu, which define the data needed by a class of documents, such as Vendor Invoices or Customer Orders.  There can be any number of definitions, each describing the data fields required for the downstream purposes of the job.  Typically, this means capturing enough data to both identify the document in the archive repository, and perform post-processing, often used to produce transaction data to integrate with an ERP system.  Definitions support both AI-generated data and user-entered or calculated data.  Many types of data are supported, including text, notes, dates, numbers, selectors, and tables.  Extensive data validation is possible, used to ensure only valid data is passed to the post processing and used for identification and indexing in the archive repository.
 
An inbound source is linked to a definition so that all documents that arrive in that source are initially processed by the assigned definition.
 
A primary goal of a definition is to obtain valid data automatically through proper instructions to the AI engine.  A fully validated document can pass from inbound source to archive and post-processing without human intervention, producing a fully automated process.
 

•IDP Documents, where IDP users can manage documents that require manual handling, such as those with validation errors.  Users can edit data as needed, reprocess through different AI models if desired, delete unwanted or mistaken documents, or transfer documents to a different inbound source for different processing.  Once data bases validation though editing, the user can perform the transfer to archive repository and post-processing.