Large screen with data and visualizations

This article was originally published in Legaltech News.

‘Defensibility’ is more than just e-discovery. Predictive analytics can take the guesswork out of crucial decisions—especially ones that impact the bottom line.

E-discovery is a big part of litigation—due to both the sheer document volume as well as the difficulty of ensuring only the right documents get produced. Yet under court rules, litigants must be able to demonstrate to the judge that their methodology for storing, categorizing, retrieving, producing, and destroying documents is “defensible.” They must be able to prove that it is sound and trustworthy.

This concept of “defensibility” is useful beyond the realm of e-discovery, though. Corporate legal departments (CLDs) are sitting on large stashes of raw data that could be put to use internally—for everything from choosing an outside law firm to estimating settlement amounts.

Failing to use this data to make predictions will soon no longer be defensible in the eyes of those who expect CLDs to run like businesses. Estimations based on “gut” or “experience” will be increasingly viewed as lackadaisical, methodologically unsound, and untrustworthy. Law departments who continue to rely on gut instinct alone will be viewed by their colleagues as out of touch.

Defend Your Decisions with Data

Some CLDs feel like they are data-driven because they look at historical data to aid humans in predicting things like life-of-case budgets. But AI has the ability to analyze far more information than a human and uncover patterns humans could never have discerned. Best of all, machine learning gets smarter every day and offers better and better recommendations.

Predictive analytics lets CLDs delve deep into historical and current data, along with industry-wide statistics, to make important recommendations regarding which outside law firm is best-suited for a particular matter, how long it might take to get that matter resolved, and what staff is necessary for success. Previously, attorneys had to make such decisions based on their instincts or through a labor-intensive process using Excel spreadsheets, but no longer. According to the 2018 Altman Weil Chief Legal Officer survey, 9.1% of CLDs are already using AI to predict outcomes in at least some portion of their legal matters. Going forward, that number will expand and become the new normal.

The Nuts and Bolts of Predictive Analytics

Defensible decision-making in any CLD starts with data. Predictive modeling is dependent on the amount and type of data you share with the system; more data means smarter recommendations. A data strategy must be put into place and executed or success will be only partial.

To create this strategy, ask yourself: what predictions do we want to make? What information do we need to make those predictions? Do we have that information stored somewhere—and in what format? For example, let’s look at how to predict litigation settlements and other outcomes like time to close. While insurance claims departments and some of the more sophisticated corporate law departments record outcomes, many others still do not. They also often do not record many of the other data points that would feed into predicting outcomes, like venue, judge, the claims listed in the complaint, opposing counsel, etc.

Making sure all of the relevant data is recorded is critical to allowing machine learning to realize its promise and ability to continually scan large data sets and add new data points over time. Once a case is resolved, for example, everything from cycle times to expenditures can be fed into the system and used for future decision-making.

Predictive Analytics at Work

Selecting the right outside counsel for a case is complex. Far too often, legal teams and claims managers deny or are unaware of its complexity and choose law firm partners based on “recency bias” (the most recent one they worked with) or without a detached analysis of the pros and cons.

Of course, selecting counsel is just the beginning of the process. Other questions linger: How much should we settle for, if we settle at all? What’s our budget? How quickly should we expect to get out of this case? Once again, predictive analytics can turn these decisions from instinct-driven to data-driven. This is particularly useful for newer associates with less experience.

The Bottom Line

Predictive analytics combines data, AI, and machine learning—a combination that can have a tangible impact on the costs and workflow of CLDs. In fact, such cutting-edge technologies will soon be the new normal, as they hold tremendous power to help attorneys sift through large amounts of information and make better decisions, from validating firm selection to identifying red flags in a budget.

Not only is this superior to gut instinct, it is also more likely to be viewed by business people as a defensible, evidence-based methodology they can trust.


About The Author

Nathan Cemenska

Nathan Cemenska, JD/MBA, is the Director of Legal Operations and Industry Insights at Wolters Kluwer's ELM Solutions. He previously worked in management consultancy helping GCs improve law department performance and has prior experience as a legal operations business analyst.

In past lives, Nathan owned and operated a small law firm and wrote two books about election law. He holds degrees from Northwestern University, Ohio State University, and Cleveland State University.