Amid a flurry of activity this summer, UK-based legal bot, LawBot, has rebranded its overall name to ‘Elexirr‘ and challenged the legal community to a battle to prove the accuracy of its case prediction software.
As previously covered in Artificial Lawyer, LawBot/Elexirr has announced that it will offer law firms two new services in addition to its advice bot. The first will be to help match potential clients with law firms, primarily by using data from users who come to the legal bot seeking advice and information. I.e. LawBot becomes a BD tool that the law firms can then use to confirm potential client leads. The matching system also takes into account the type of law firm and factors such as its location to get the best match.
The second offering is aimed at corporate clients and forms an analytics package, which the LawBot team describes as: ‘legal analytics for private companies in, or considering, litigation’. In short it estimates the likelihood that a particular type of dispute will be won or lost.
The company states: ‘Our analysis considers all features of a case, including the facts, profile of the parties to the dispute, and legal problem. Our algorithm compares the supplied information with a large database of real legal problems with known outcomes (e. g. succesful insurance claims) and trains a model on the most relevant documents to make a prediction. Elexirr’s accuracy comes in at over 71% using a K-fold cross validation with a K-factor of 10. From this information, we are able to provide a tailored analysis of your company’s chance of winning the claim.’
It is this second application that LawBot wishes to test out against the predictive capabilities of law firms in an open challenge. It is hoped this challenge can take place at some point in the next few months.
LawBot co-founders, Ludwig Bull and Rebecca Agliolo (pictured above) said: ‘To demonstrate the utility of our system, we are organising a competition between real lawyers and our predictive algorithm. An independent panel will supply factual and legal descriptions of real legal problems that have a known outcome (e. g. an already decided case). Both the human and the machine team will then have time to consider the problems carefully and make a prediction for each one. Whichever side makes the most correct predictions, wins.’
The announcement marks a busy period for the former Cambridge law students, who have also announced a plan to launch their own digital currency to fund the growth of their company later this year, which Artificial Lawyer covered previously.