The Affinda solution enables users to create data look-ups and matching logic to enhance the data extraction process. With these lookups, you can choose to:
- Validate the data extracted
- Populate a data field
- Overwrite a model prediction
The following are common uses of the data matching logic, however, the matching and lookup functionality is built to be general in nature and the logic can be applied across a range of use cases.
A customer may have a Purchase Order database and wants to assert that the total amount due is equal to that in the purchase order. Using Affinda, we can:
- Extract the Purchase Order from the document
- Match the Purchase Order number extracted by Affinda to a Purchase Order number in the database
- If the corresponding 'invoice total' value in the Purchase Order database matches the 'invoice total' that Affinda extracted, this value would be auto-validated.
A customer may have a Supplier database and wants to assert that the supplier found in the invoice exists within their master list. Using Affinda, we can:
- Extract the Supplier Name from the document
- Apply fuzzy string matching to match against a supplier name in the database
- Populate Supplier Code with the corresponding in the supplier database
Updated 5 months ago