Ephesoft Advanced Capture Rebranding
At the Innovate conference, Ephesoft announced that starting with version 4.1, their Enterprise Advanced Capture product would be rebranded as Ephesoft Transact while Ephesoft Universe would be rebranded as Insight. This makes sense in light of what each is used for. Universe provides Insight into you data collected from your documents. Enterprise Advance Capture tends to be more Transactional; data from a document is extracted for a particular transaction, such as processing a mortgage, invoice, medical record, or new employee.
Transact Version 4.1 officially released (November 17, 2016)
Among other features, the newly branded version of Transact touts improved document classification and data extraction.
LAGUNA HILLS, CA–(Marketwired – November 17, 2016) – Ephesoft Inc., creators of the industry-leading Smart Capture® advanced document capture platform, today announced the release of Ephesoft Transact 4.1 (previously named Ephesoft Enterprise), the company’s cloud-based platform for automating document driven processes. Ephesoft Transact was previewed for customers and business partners at the company’s recent technology conference, Innovate, held in Irvine, California.
Key new features in the release include dynamic document classification, ERP integrations with progressive line item matching between purchase orders and invoices, a new secure web scanning service that eliminates applets while supporting https protocol, and advanced machine learning for document classification and content extraction.
“Our philosophy in developing Ephesoft Transact is to focus on our core premise of delivering an easy to use, intelligent document capture platform so more organizations can benefit from the technology,” explained Ike Kavas, Ephesoft’s Chief Technology Officer. “The hallmark of Ephesoft Smart Capture technology has been the ability to identify documents with few samples and extract data fields regardless of format. With machine learning introduced in Ephesoft Transact 4.1, we make both these processes easier for administrators and operators.”
How Transact Classification Works
The document classification is now backed by Apache Spark in addition to leveraging the traditional OCR based document classification. In case you didn’t know, when Transact (Enterprise) classify documents, it looks at the words on a newly processed page and compares that information to other pages it has been trained to recognize. It assigns a score to matching known pages. The document assembler then looks at those scores to decide what pages make up a document and where the document breaks are.
The addition of Apache Spark along with the concept of perpetual learning means that as users correct the results of document classification, Transact will leverage those results when doing future comparisons and scoring. In theory you can now implement Ephesoft Transact by defining document types with no document training and let the users train the system in production.