Anomaly detection leverages Inscribe's database of documents to identify discrepancies between your submitted document and documents of the same type we've seen in the past. Anomalies are identified using the 'invisible' metadata associated with the document, as well as details visible to the human eye.

Anomaly detection can identify issues with the fonts present in a document, as well as the formatting of key details such as names, addresses, dates and amounts. While there is no one-size-fits-all for formatting these details, Inscribe is able to anticipate the expected format for a given type of document.

When anomalies are found relating to details visible on the document, the corresponding regions on the document are highlighted. You can cross-check the expected format by manually examining the ‘Similar documents’ section.

False positives may occur when a service provider changes the formatting of their documents. However, our system will progressively assimilate new data and refresh its database, solving these issues over time. False positives can also occur if the submitted document has been formed by merging multiple different documents together, and details introduced from the merge are identified as anomalous.