Law has a Data Problem
Agnes Legal is built on the partnership between John Snyder, a top New York City litigator, and John Macartney, a data scientist and pioneer in the field of concept clustering. Together they are solving the problem of eDiscovery
eDiscovery cost Americans $42 billion dollars in 2017 and will cost more in 2018. These skyrocketing costs, often in the millions of dollars per lawsuit, create an environment where only large corporations and the wealthy have access to justice. These problems exist because we are still using 20th century technology to deal with a 21st century data problem. Inefficiency has a human cost, both in unequal justice, and in burnout from mindless, repetitive work. AI assisted data technology allows humans to cope with mountains of information quickly and simply without sacrificing transparency, accountability, or human discretion.
Our technology understands the content of communications, whether they be text, audio, or video. It extracts ideas from the words, and knows what is relevant to the user– not just keywords, but their meaning and the ideas behind them, as well as related concepts. It displays the data in a clear, meaningful way, allowing users to track ideas through vast amounts of information.
At the beginning of the review process, the user outlines the elements of the case, and whatever basic information they consider relevant (usually names, dates, and keywords). Based upon this initial information, the algorithm extracts a series of concepts from the data and displays them to the user.
The user responds to each displayed concept by clicking left (not important), right (important), or up (hot). This process is called Concept Sifting, a unique process invented by Agnes Intelligence which shapes the data so that the most relevant documents bubble up to the top. Each action is recorded and factored into our algorithm, so that the technology learns iteratively as the user evaluates more and more documents.
Because there are a finite number of important documents in a case, the user will see fewer and fewer relevant ones as the review progresses. The Snyder Score is a metric to show how far the review has progressed. When the Snyder Score reaches 100, the user has seen all the relevant documents.
During the review, the best documents are automatically tagged and sorted into witness folders, used to populate a detailed timeline, and mapped to the elements of the various causes of action. At any time, the user has the option to conduct additional searches.
Our user interface is simple– it provides what a trial lawyer actually needs: a timeline, witness folders, and an outline of proof. No incomprehensible menus. No complicated user interface. No straining your eyes to click tiny boxes. Just the relevant documents. First.