Daniel Katz is a scientist, technologist and law professor who applies an innovative polytechnic approach to teaching law, meshing litigation and transactional knowledge with emerging software and other efficiency-enhancing technologies to help create lawyers for today's challenging legal job market. Both his scholarship and teaching integrate science, technology, engineering, and mathematics. Professor Katz's forward-thinking ideas helped to earn him acknowledgement among the Fastcase 50, an award which "recognizes 50 of the smartest, most courageous innovators, techies, visionaries, and leaders in the law."

Follow Daniel Katz’s Computational Legal account on Twitter @computational

Professor Daniel Katz provided a great finish for ELM16 with his keynote entitled Measure Twice, Cut Once. Solving the Legal Profession's Biggest Problems Together. With a focus on how the legal industry can use metrics and data to thrive in changing times, Daniel’s expertise proved very thought-provoking for the conference attendees.

He began by explaining the fundamental economics of the legal industry. At the most basic level, the value prop of lawyers is that they help to navigate complexity and manage enterprise legal risk. We are at the end of a “golden age” of profitability for law firms. As legal complexity grew in modern business, we kept throwing more and more labor at the problem to get it solved. And with that labor charging by the hour, legal spend went up and up.

Since 1980, total legal expenditures relative to the United States GDP went up somewhere between 2x and 4x. But we have reached a point where corporations can’t continue to increase the amount of money spent on legal work. They need to find tools that will help them solve legal problems by some means other than more lawyers working for more hours.

Like many areas of commerce have before, the legal space is undergoing a shift from artisanal to industrialized delivery. The challenge is to retain the artisanal aspects that add value to law and industrialize the rest. It’s analogous to something we see with pizza, which Daniel admitted is a favorite treat. The independent, corner pizza restaurant may have the most delicious pizza, but they aren’t scalable and can’t get their product out beyond the local area. Meanwhile, Domino’s may not have the best pizza, but they can get it to a huge percentage of the population very quickly. The question is how the legal industry can achieve the needed scalability without sacrificing quality.

Daniel believes that data analytics provides an answer. With the right data, it is possible to make two kinds of predictions:

  • Procedural predictions about things like case cost and duration
  • Substantive predictions about case outcomes

Procedural predictions aren’t all that difficult to arrive at and they provide useful information about averages as well as more unusual cases that appear as outliers in the data. But the “holy grail” is the substantive predictions that allow corporations to minimize risk and exposure by revealing the most likely outcomes of cases. When these substantive predictions are granular enough, they can even enable mapping of particular contract terms to financial results so that staff can concentrate their negotiations on the most important aspects of deals.

The most effective predictions, Daniel has found, do not rely on a single source. Ensembles of expert predictions, crowdsourced predictions, and predictive algorithms yield the most reliable results. Machine learning and AI in the Cloud are likely to make the algorithmic aspects of these ensemble predictions cheaper and more accessible, helping to move Legal toward that partially-industrialized goal.

As Daniel and others in the field continue to work on understanding how data can help to manage risk, increasing efficiency and profitability, more products and services will be introduced to focus on helping corporations and law firms improve ROI and meet financial goals.

Much of law is now practiced as an “art,” but Daniel explains that more aspects of legal work should be moved into the “science” column. “I hear all the time ‘You can’t really predict [x].’ Really? You can’t? Or you just don’t know how to?”