Today, Seal Software announced the launch of its version 4.1 release, with significant new features and capabilities that address three core areas of the platform: The interactive visualization of contractual information, the accuracy of that information, and the rapidity in which the platform extracts data from contract documents.
A Picture Paints a Thousand Words
Firstly, how users receive and interact with information, addresses the challenges systems have in presenting “important data” output to users. How do you extract information from 10,000s to 100,000s of contracts, and make it so business users – not legal professionals or highly trained analysts – can build and consume it in a way that makes sense to them? You incorporate a powerful interactive data visualization and reporting engine that allows users to create custom reports and dashboards and present different visualizations, thus providing the user to the ability to explore these visualizations through a single “pane of glass.”
With V4.1, Seal has integrated the industry-leading data visualization and reporting engine from Qlik software. Qlik has recently been announced as a Leader in Gartner’s MQ for Business Intelligence and Analytics platforms, and comes out-of-the-box and integrated with the Seal platform. This means users of Seal are never buried in data, but get it presented clearly, intuitively, and in a way they can make more effective business decisions about future contract expiration, governing law geo mapping, obligations and incentives, and optimize their commercial and legal relationships.
Under the Hood
The second significant new capability in V4.1 is an entirely new machine learning (ML) framework. The machine learning within Seal enables the teaching and user defined analytics aspects of the platform, and in conjunction with our Natural Language Processing, is how Seal has significantly raised the bar for accuracy and speed in our industry.
With V4.1, Seal has added several new algorithms in a “Deep Learning” methodology to dramatically increase overall accuracy and speed with our contract analytics. In previous releases of the platform, we used MaxEnt and MaxEnt with CRF, but with V4.1 we’ve added several new models including Support Vector Machines, Word2Vec with CBOW, and K-NN.
To optimize the output of the new models, Seal has developed its own ensemble algorithm to orchestrate the layers of algorithms being used together to identify contract data. The results will set the bar – again – in the efficiency and precision of extracting user-defined terms and provisions from contracts.
Enterprise Performance & Scalability
In the aspects of speed and scalability, our Development team in Gothenburg didn’t want to stop with the new ML framework. Some of our customers are engaging Seal with several hundreds of thousands of contracts to ingest and review and ask new questions of, so the team increased the speed that Seal can ingest contract documents by 5x. Recent benchmarking tests have shown the capacity to ingest documents at a rate of over 1.9 million contracts per day, which shortens the time-to-information for all our customers, but especially those with very large quantities of contracts to process.
The product team was also busy working on a new version of our API (V 2.0) for system extensibility, and an integration with DocuSign that allows Seal to quickly report on which contracts have signatures, and which ones are unsigned.
Seal V4.1 will help all of our current and future customers with increased speed and scale, improved accuracy with our new ML framework, and the new data visualization and reporting engine that will give business users the contract insight and intelligence they need, just the way they want it, in a timely fashion.