Ephesoft Transact 2022.1 Breaks Barriers with New Semantik AI Engine

Ephesoft’s intelligent document processing platform now has the ability to extract data from unknown documents out-of-the-box, accelerate known document design by up to 90% and more

Ephesoft Inc., a global leader in intelligent document processing, automation and data enrichment solutions announced the new release of its IDP platform, Ephesoft Transact 2022.1. The new version of Transact features cutting-edge AI, computer vision and deep learning neural network technology, named Semantik AI Engine™, which powers data extraction from both known and unknown document types.

“The future of intelligent document automation lies within forming new ways of using AI to deliver the right data to our customers,” said Ike Kavas, founder and CEO at Ephesoft. “The ability to extract data from unknown document types means that users can now pull data from any document with no configuration. Clients can now find and access previously unknown data as well as accelerate hyperautomation through iPaaS and business intelligence systems. This will allow customers to easily expand their investment in Transact to new use cases in the cloud and on-premise with constantly improving AI models.”

Universal Document Automation

Ephesoft’s new Semantik AI Engine in Transact enables document processing for unknown document types. This feature allows for document automation without the need to create a template, train the system or configure the solution. It can recognize any document type and identify key-value pair entities for any structured and semi-structured document type out-of-the-box.

Read More:  OpenBots Announces ‘Zero Bot License’ RPA Revolution Tour

• Process unknown documents
• Automatically detect and extract critical document data
• Data mine and opportunistically explore any data from any document without configuration
• Enable data extraction projects that would otherwise be too costly or time-prohibitive

Universal Document Automation is a new, additional approach to document extraction in Transact to serve specific use cases that include a large volume of unknown documents; its pre-trained model can be deployed both on-premises and in the cloud, and works with English, Spanish, German, French, Italian, Polish, Swedish and Dutch languages.

“As part of Ephesoft’s Early Adopter Program, I was able to test out the Semantik AI Engine built with computer vision technology. It was amazing,” said Liz Burnside, Sr. Solutions Consultant at Zia Consulting. “I uploaded batches with many different document types like shipping docs, receipts, POs and the data was extracted instantly. It was very thorough and extracted all the data that I needed in less than a minute; 10-20 header and footer fields were correctly extracted per document. The Universal Document Automation feature was able to quickly capture and extract all the information from various documents without set-up.”

“I have been testing the Universal Document Automation feature for the past few weeks and I have to say, it is going to be a game changer,” said Lisa Hill, Vice President of Technology Strategy, ArgonDigital, who also participated in the Early Adopter Program. “We work with clients in oil and gas that have shared drives (or even file cabinets) full of old well logs, completion packets and other documents that they know hold critical information, but they really have no idea what is in there. Being able to quickly extract information without any prior setup is going to be huge.”

Read More: SalesTechStar Interview With Todd Donahue, CFO At Ordergroove

Document Design Accelerator

Another innovative capability of the new Semantik AI Engine is the Document Design Accelerator, which significantly speeds up the process of building index fields and extraction rules for new, known document types. All that is needed is to supply a sample document to the Document Design Accelerator and it will analyze the content and automate the conventional process for enumerating index fields and creating extraction rules. Results from the Early Adopter Program showed that setting up new documents was expedited by up to 90%.

Our Early Adopter, Burnside continued, “And, for documents with fields I knew I wanted to configure, I tested the Document Design Accelerator feature with a form that previously took me a few hours to set up. The accelerator allowed me to do this in seconds, which will benefit customers’ hyperautomation initiatives and save costs.”

Other highlighted features in the new release include:

• iPaaS Connectors: Transact now offers two new connectors in the MuleSoft Anypoint Exchange platform and the Workato Connector Library community. The connectors enable a low-code/no-code Transact integration into thousands of systems to enable fast, seamless end-to-end automation.
• PDF XFA Flattening: Converting PDFs that use the XFA technology into images has been a longstanding challenge because of the special way information is stored in these PDFs. By expanding the folder import capabilities, PDF XFAs can now be flattened, which simplifies and expedites extracting data from these types of documents.
• Cloud OCR Plugin: Ephesoft now offers clients flexibility to call advanced cloud OCR engines directly from their Transact deployments via a new plugin. Google Cloud Vision can be used for handprint, cursive and machine print extraction through a direct connection to their existing Google Enterprise infrastructure. Additional cloud OCR engines will be added to upcoming releases.
• Improved Traceability for Extraction Rules: Users can now easily identify which rule has been applied for the extraction of each index field. The extraction rule identification will be available as part of the document processing information, allowing easier fine-tuning and maintenance for rule sets.
• Expanded Environment Support: Customers with Red Hat Enterprise Linux 8.4 and SQL Server 2019 versions are supported.