Seal Software Rolls Out Version 5.3 of Artificial Intelligence Platform for Contract Discovery and Analytics

Share Article

Latest Release Extends Rich Capabilities of Seal Platform With Introduction of ‘Active Learning’ to Quickly and Intuitively Create New Machine Learning Modules

The latest version of the Seal platform puts even more power into the hands of business users, making it possible to do more and achieve more without the need for legal or IT resources.

Seal Software, the leader in Contract Discovery and Analytics, announced today that it has released the newest version of its award-winning platform. Incorporating artificial intelligence and natural language processing, Seal’s advanced technology performs automated discovery of contracts across the enterprise and brings them into a centralized repository for extracting all of the valuable terms, provisions, and clauses buried in the text. Version 5.3 builds on Seal’s machine learning engine and analytics framework introduced in Seal V4.0 with a new process called Active Learning, which allows end users to quickly and intuitively create new machine learning (ML) models without the need for data scientists, consultants, or technical staff.

The latest version also enhances the Analyze This Now feature with a new Clause Library that can be used to swap out unapproved language with approved text while working on contracts within Microsoft Word, and includes web services to expose this functionality to any external application for insertion, tracking, and reporting on clause use. Moreover, the new version introduces out-of-the-box German-language extraction models for critical contract components, and includes a new centralized configuration capability.

“The latest version of the Seal platform puts even more power into the hands of business users, making it possible to do more and achieve more without the need for legal or IT resources,” said Kevin Gidney, CTO of Seal Software. “With Version 5.3, our customers continue to gain ever greater value from Seal to drive new revenue and gain competitive advantage over business rivals.”

Active Learning

The key to accuracy and scalability with any system which uses machine learning is the ability to train the system to ensure it knows what to look for and how to analyze data. Introduced in Version 5.0, UDML extended Seal’s machine learning framework by allowing end users to show the system positive and negative examples of the types of contract text to search for from across their entire portfolio as they create custom extraction policies.

With Version 5.3, Seal has introduced Active Learning, a unique feature to the market which significantly speeds up learning and creation of machine learning models. Instead of training the system by finding example text across various documents, users are now presented with a document preview that highlights a particular set of text for evaluation. Users can then indicate whether or not the text is appropriate for the specific extraction policy.

For example, users can intuitively select “Yes, this is a good example,” or “No, it is not a good example for this model,” or skip to the next example. End users can create digital applications with the combination of the Active Learning, information normalization, and presentation framework, reducing or removing the requirement to utilize specialized personnel.

Gidney continued, “Active Learning is the next step in Seal’s path to ‘one-shot’ learning. By combining all of our ML methods with a set of proprietary processes and function, the Seal platform is now able to quickly and effectively generate a highly performant model with minimal review and examples.”

Additional Features

Version 5.3 also extends the Analyze This Now feature introduced in Version 5.0. This allows end users to use Seal while working on contracts from within Microsoft Word and have the system inform them if their clauses or provisions are using approved language or not, and web services are available for other application frameworks to integrate this capability. With Version 5.3, if Seal shows an end user that there is a clause or clauses that do not use approved language, Seal’s new Clause Library can be used to swap out the unapproved language with approved language.

This means that users don’t have to ask their legal resources, either in-house counsel or third-party counsel, to evaluate the language and provide them with appropriate text. The result is a dramatic cost and time savings as legal teams no longer have to find, manage, and replace non-standard language with standard and approved language.

Additionally, the latest version of Seal now includes of out-of-the-box German-language extraction models for critical contract components such as assignability, auto-renewals, contracting parties, governing law, limitations of liabilities and more.

About Seal Software

Seal Software is the leading provider of contract discovery, data extraction, and analytics. With Seal’s machine learning and natural language processing technologies, companies can find contracts of any file type across their networks, quickly understand what risks or opportunities are hidden in their contracts, and place them in a centralized repository. Based in San Francisco, Seal empowers enterprises around the world to maximize revenue opportunities, reduce costs, and mitigate risks associated with contractual documents, systems, and processes. For more information, visit Seal Software at

Share article on social media or email:

View article via:

Pdf Print

Contact Author

Brigida Bergkamp
Spark for Seal Software
+1 2537773543
Email >
Visit website