OpenText Decisiv brings Predictive Research to Enterprise Search

AI-enhanced search engine now leverages supervised machine learning to find useful content

It’s been nearly a decade since Recommind (now OpenText) first pioneered predictive coding for legal document review with OpenText™ Axcelerate™. The advent of supervised machine learning revolutionized eDiscovery with the simplest of principles: if those documents are of interest to you, these probably will be also.

Now, all users searching for content across enterprises can benefit from this same principle — and the same proprietary artificial intelligence technology — with Predictive Research in OpenText™ Decisiv™ 8.2, released in advance of OpenText Enterprise World 2018.

Decisiv augments enterprise search with the power of artificial intelligence (AI). Decisiv’s unsupervised machine learning provides sophisticated, conceptual analysis of unstructured content to help users find what they’re looking for, faster. Decisiv helps legal teams, government workers and business professionals identify useful content, and people with specific expertise, by instantly searching across a wide range of enterprise sources, even when those searching are not sure what terms are likely to yield the best results.

For example, when a Decisiv user types “asia,” the system instantly and automatically retrieves documents that (1) contain the word Asia and/or (2) are conceptually related to Asia, regardless of whether they contain the word “asia.” Decisiv not only casts a wider net, it delivers and prioritizes results based on more sophisticated relevancy analysis than simple keywords can provide.

“Asia not Asia” search yields documents conceptually related to Asia that nevertheless don’t contain the word “Asia”

With Predictive Research, Decisiv now adds supervised machine learning to its functionality: the system learns on a continuous basis from human decision-making. As users “pin” useful documents to their Research view for quick access, Decisiv automatically learns what they are interested in, and automatically suggests relevant documents.

Importantly, Predictive Research requires no change in how researchers operate. A single click will pin a document to the Research view, a convenient space for ready access. The Research view operates like a virtual desktop, easily collecting key content without requiring downloads or making additional copies of files. Based on sophisticated document models generated from those pinned results, Decisiv automatically – virtually instantly – suggests additional documents likely to be useful to the researcher. A single click on a suggested document enables the user to approve or reject the suggestion, automatically refining the system’s training and updating the suggestions.

Research view shows documents “pinned” by the user and “Recommended” documents from Predictive Research

Using Predictive Research also improves overall search results across the enterprise. The more a document has been pinned by users, the more likely it is to be prominently featured in Decisiv search results. This valuable contribution of tribal knowledge is leveraged across the organization to help the search engine automatically pinpoint valuable content.

As it helps deliver what users are looking for, even without their asking, Predictive Research is one of those advances that seems likely to be considered indispensable in the not-too-distant future.

Join us at Enterprise World to learn more about Predictive Research and other AI use cases for legal teams and business users across your enterprise. Register today!

 

Hal Marcus

Hal is the Director of Product Marketing for OpenText Discovery & Security. A licensed attorney, Hal practiced as a Wall Street litigator before commencing a career in tech that now spans over 20 years. He writes about artificial intelligence—anyone named Hal is bound to be interested in AI—and related technologies in the realms of eDiscovery, cybersecurity, compliance, and information governance.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Close