Outcome Prediction in the Practice of Law

Business forecasters typically use time-series models to predict future demands, the forecasts informing management decision making and guiding organizational planning. But this type of forecasting is merely a subset of the broader field of predictive analytics, models used by data scientists in all manner of applications, including credit approvals, fraud detection, product-purchase and music-listening recommendations, and even the real-time decisions made by self-driving vehicles. The practice of law requires decisions that must be based on predictions of future legal outcomes, and data scientists are now developing forecasting methods to support the process. In this article, Mark Osbeck and Mike Gilliland first examine the traditional tools lawyers employ along with the limitations that prevent these tools from consistently delivering accurate predictions. They then describe how new data-science approaches, including AI, are starting to alter the way law firms operate.