Bostinno
How ArbiLex uses AI to predict the odds of legal cases
By: Jordan Frias
After being burned by her former best friend, ArbiLex founder and CEO Isabel Yang was determined to become a lawyer.
As Yang tells it, her friend cut her out of the profit share in a solar power venture they started together in South Africa. Unsure of how to handle a legal dispute — especially in a country nearly 8,000 miles from Boston — she decided to learn everything she could about international arbitration.
“I think what that story really left with me was there are uncertainties that are unique to doing business cross-border,” Yang said. “Not knowing the odds that I’m facing in a case… It’s a very helpless feeling.”
Now, she’s made international arbitration the focus of her startup, ArbiLex. Yang combined her statistical knowhow from her days at Oxford with a newly earned Harvard law degree, then recruited three other people to begin working on a tech platform that would cater to the legal industry. Working out of the Harvard i-lab, the team created the subscription-based “Arbitrator Analytics Platform,” an AI software that can determine things like how much a case will cost and the likelihood of who might win.
“Think of us as the Moneyballers for lawyers,” Yang said. “All ArbiLex is doing is we take variables and features of a case that are indicative of a winning case, and we quantify the impact of those features on the chance of success of a case.”
Yang said pattern-spotting based on case law and jurisprudence allows ArbiLex to find the characteristics of a winning case while using AI and predictive analytics. She thinks parties in high-stakes legal disputes could benefit from the predictive analytics her startup provides.
ArbiLex customers tend to be large law firms, but they also include bankers, Yang said. ArbiLex’s data-driven predictions allow bankers to decide which cases to back. Those banks then get a percentage of the law firm’s winnings, should the firm prevail in court.
“A lot of the demand for our data analytics actually [comes] from the litigation [financers], because they are looking to bet on certain sides of a dispute,” Yang said. “They need to understand the risk and uncertainties, and our sole job is to quantify uncertainties.”
Lawyers also see value in the platform, she said, because they can let bankers know that they have a strong case and a high chance of success based on ArbiLex’s research.
Although Yang had a personal motivation to focus on international arbitration, she is not completely new to this world. She became fascinated with the legal aspects of international policymaking through her prior work for the World Bank and the United Nations, and as a senior economist for the Minister of Finance in South Africa. Those experiences taught her that decision science could be applied to the legal services industry, she said.
Three years ago, ArbiLex got the backing of several angel investors, including Kensho Technologies founder and CEO Daniel Nadler. Yang said his backing gave her company a lot of credibility.
“I think in his own way, he’s doing what I’m doing for legal services for financial services,” Yang said.
Now, Yang’s goal is to get ArbiLex into the asset management business and fund some of the cases it is already reviewing. That will require more fundraising, she said. Yang also plans to set up physical offices in Boston and New York sometime in 2021.
With international arbitration cases among the fastest-growing practice areas for global law firms, Yang hopes potential clients take note of how the ArbiLex platform could drive efficiency.
“Over time, as we scale up, I think the impact we’re going to make in [the] legal services industry is that the pricing of legal assets will become a lot more data-driven,” Yang said. “Really using data and technology liberates us from a lot of the unnecessary groundwork that inevitably goes into the process by using solely a manual process.”
