By Julius Melnitzer | March 22, 2021
Lenczner Slaght Royce Smith Griffin LLP is incorporating homegrown data-driven decision-making as a key part of the firm’s litigation strategy.
The prominent litigation boutique is certainly not the first firm in the country to use machine-learning predictive outcomes – but they appear to be the first to have embraced a comprehensive data harnessing program founded on proprietary databases.
“We created our databases manually, and we used a fairly common machine language for our predictive analytics software,” says Paul-Erik Veel, who leads the program, which the Toronto-based firm calls Lenczner Slaght Data-Driven Decisions. “But because lawyers will tell you that 200 things matter in a case, we had to focus on collecting only the information that had value in predicting outcomes – and that required some long conversations about what actually matters.”
The initiative came about because clients have been pressing for more objective certainty in the advice they get.
“We expect doctors to know the probability of outcomes when they’re doing things, but we’ve never held lawyers to the same standard,” Veel observes. “Sophisticated clients now want that empirical data, but lawyers are not used to working with and communicating it to clients.”
In some cases, lawyers can’t even give their clients the most basic statistical information.
“I’m endlessly frustrated, for example, that there is no answer as to the proportion of class actions that achieve certification in Ontario,” Veel says.
The Data-Driven Decisions initiative has three prongs:
- Harnessing available technology and products that make use of data analytics;
- Remaining engaged with pioneering empirical research on litigation and advocacy; and
- Developing in-house proprietary data sets and analytics.
So far, Lenczner Slaght has three projects in the works:
- The Supreme Court of Canada Leave Applications Project, which uses a proprietary dataset containing information on 1,500 leave applications. The project employs a machine learning model that predicts both the likelihood of getting leave and how long it will take for the court to release a decision;
- The Commercial List Project builds on the firm’s proprietary database tracking 40 characteristics of all Commercial List decisions since 2019. It is the foundation for Lenczner’s annual Commercial List Year in Review report; and
- The Competition Tribunal Project database features every case filed at the Tribunal since the late 1980s. It analyzes more than 70 variables to provide objective perspectives on the risks and potential outcomes of enforcement actions by the Competition Bureau.
The program has already engendered significant benefits. By way of example, a client called the firm seeking the enforcement history of a particular Competition Act provision.
“In the past, we’d have an articling student scour the website and come up with a list of cases, which entailed significant time and cost to the client,” Veel says. “In this case, we ran a query and had an accurate answer in five minutes at absolutely no cost to the client.”
Toronto-based Blue J Legal Inc. is a pioneer in using machine learning software to predict outcomes of specific scenarios. The company claims that its software has 90 percent accuracy on average.
“Our mission is to bring absolute clarity to the law and on demand,” says Benjamin Alarie, the Osler Chair in Business Law at the University of Toronto as well as Blue J’s co-founder and CEO. “The path to legal singularity may take decades – maybe centuries – to accomplish, but the endpoint is a clear and immediate sense of what the law requires.”
Decades away notwithstanding, it seems there’s no shortage of believers now: over 200 law firms in Canada and the U.S. have signed up for Blue J’s predictive tax and employment law products.
Julius Melnitzer is a Toronto-based legal affairs writer, ghostwriter, writing coach and media trainer. Readers can reach him at [email protected] or https://legalwriter.net/contact.