At a Legaltech of long ago, I was given a demonstration of the discovery software of the company then called eDiscovery Tools. I asked if the software included a particular feature, and the demonstrator thought for a while and said “We don’t have that yet. Ask me again in two weeks”. I was rather impressed by this reply from what was then a small software company, with the ability to assess a suggestion and implement it quickly.
That company, now called EDT, is no longer so small. As founder and CEO Jo Sherman says in this video, the company has grown considerably over the last two years. I asked Jo Sherman to tell me what EDT is working on now.
The most exciting projects, Jo Sherman said, are those going on in EDT’s AI lab in Israel and elsewhere, much of it in collaboration with clients. She gave as an example EDT’s work on sentiment analysis and profanity detection. When people are under stress or “doing something they should not be doing”, they tend to use more profanity in their communications. EDT has been working on following up certain terminology in order to get to the “good stuff” in a dataset.
EDT has been working as well on how machine learning tools are used – technology-assisted review and continuous active learning are by now well-established, and EDT has been thinking about how to re-engineer how the work is done. One project has been on similarity analysis, allowing quick and effective searches for documents similar to those already identified as critical, and on effective ways to apply unsupervised learning to large datasets.
The emphasis, Jo Sherman said, is less on documents and more on conversational communication and on interactions between the key people in a case.