Companies lose billions of dollars in litigation suits worldwide every year. Quite often, they are taken by surprise when the court order comes in and are helpless in response.
Step in Intraspexion.
Intraspexion is a unique company in the legal world. The organization uses predictive analytics and deep learning to identify the risk of a client being sued. Using it’s patented software system, it aims to predict and prevent potential legal suits.
So how does the system actually work? The company’s deep learning model runs through the emails throughout the enterprise to identify potential risks while the communication is still internal. The software involves mining and leveraging classification of the existing data to train the deep learning model. Following that, the algorithm is run on all the internal communication (emails). This in turn generates a score that helps the risk team to notify the concerned parties that a potential law suit could be on their hands soon.
The company claims to preserve the client’s net profits by an average of $350,000 to $450,000 per case.
Since some of the software is patented, the technology behind it is a secret. However, on their site they have revealed that parts of the software have been developed using GloVe and Google’s TensorFlow.
Take a look at a demo of the Intraspexion system at work:
Deep learning has penetrated into the legal landscape as well. Given the extensive amount of data lawyers have to sift through on a daily basis, machine learning has huge potential here. Within litigation, research and discovery are two hot fields that are already seeing ML penetration. From companies like DISCO (using predictive analytics for e-discovery) to Ravel Law (it gives lawyers deeper insights into legal data), companies are waking up to the power of machine learning.
In that sense, Intraspexion is as unique as it is transcendent. Their software has the capability of saving millions of dollars for firms and sets the precedent for others to follow.
Lorem ipsum dolor sit amet, consectetur adipiscing elit,