Overview

Azolve Extract is the method of apply automated extraction of key terms from a document. For example, if a Supplier Invoice is scanned, Azolve Extract will review this image and look to extract all the key data from the document without human keying being necessary.

Azolve Extract provides varying methods of achieving document extraction.

Template

For document types, templates can be created which are then applied to documents as they arrive into the system. These templates can be unique to specific documents, so the first stage is to determine which template should be applied.

Once a template is identified as being eligible, the template is applied which informs the application how to extract the information needed. For example, if a batch of supplier invoices is scanned, the system will check each invoice to determine which templates relates to that supplier. Once the supplier template is identified, Azolve Extract will proceed to extract the data specific to that supplier.

Zoning v Term Searching

A template can use a sample image. Areas of this image can be zoned to identify where data is expected to be located from. When another document of that type arrives, the zoned template is used to super-impose the new image and to extract the zoned areas.

Sometimes zoning is not suitable, so another technique is the use of term searching. Because Azolve Index has stored all terms within its index, Azolve Extract is able to use this to search for a term and extract data around that term – e.g. search for the word Invoice Date and extract the date that is linked to that term.

In the context of supplier invoicing, this would allow the extraction of key fields like Invoice Number, Invoice Date, Gross Total and so forth.