Cognizer provides 3 business-critical solutions

Cortex is one of the largest and most sophisticated deep learning models ever created for legal contracts. Pre-trained on millions of contracts and then fine tuned on hundreds of thousands of annotated Terms and Clauses, Cortex provides an average accuracy of over 95%.

Genius Application Suite

Genius is an application that supports Web, Mac, and PC.  Leveraging multiple Large Language Models, it is design to support business executives and legal experts with critical information from their contracts.  Users can have a conversation with their contracts and see a detailed explanation of how the AI determined the answer.  Genius provides analytics, insights, and a detailed understanding of your contracts.

Genius’s X-Ray viewer allow the user to quickly see the important terms and clauses in your contracts in a tree view next to the text of the document.  Select an item in the tree view and the viewer will automatically highlight the text in the document.  It will show the definition and each occurrence of the term in the contract.

Instant Contract Analyzer

Quickly see the important Terms and Clauses in your contracts, as well as the definition of each Term, and all of its uses in the contract. View Word and PDF Files and then immediately understand the important details of the contract

Cortex OEM Platform

 

The Cortex Platform is designed for Legal Software vendors to embed into their solution. Cortex supports up to 100 contract types and 1,000 terms/clauses. It supports the identification and extraction of these clauses with more than 95% accuracy.

In addition, the Cortex Platform provides Retrospective Learning.  Using its Neural Vector Engine provides Continuous Learning.  The system can identify previously unknown clauses and label them, as well as fin-tune the definition of existing clauses.

Finally, Cortex leverages several Pipelines to process inbound contracts.  There is a real-time pipeline for high-speed ingestion, a “dream mode” pipeline for offline processing, and a conversation pipeline for language processing and prompt engineering.  All of these pipelines provide customizable pipeline segments and order configuration.  The pipelines also support calling external classes, allowing for a “bring your own model” capability.