Smart field extraction using Athenea (OpenAI)

Athento allows the intelligent extraction of fields using Athenea (OpenAI), for this, the path to the custom extraction file must be specified in the “metadata definition” of the field to extract, which in this case is “athentose_openai.metadata_extraction”:

It is necessary to run the text extraction/OCR operations on the document beforehand for intelligent extraction to work.

For the extraction to work correctly, the AI will consider the following to build its context and determine what should be extracted in each field:

• Form: it will look at the label of the form and its description (if it has one).
• Field: It will look at the field label and, most importantly, the help text. Defining the help text correctly will help the extraction to be more accurate. For example: