10X Genomics Inc. has been valued at more than $1 billion, but a far less heralded start-up has some advice to help the Silicon Valley unicorn with a $24 million patent problem.

Less-than-a-year-old Ex Parte uses court data and a proprietary ranking system based on real-life lawyers’ experience and past results to predict who will win appeals in patent cases at the Federal Circuit.

That court is where 10X Genomics is likely headed after a Delaware jury said the biotech company owed a competitor $24 million for patent infringement and a judge ruled the company will have to stop selling certain gene analysis devices.

Ex Parte has some bad news for 10X Genomics: Its algorithm gives the company as low as a 9% chance of winning on appeal. But, by hiring one of the top lawyers Ex Parte suggests, that figure can jump to as high as 25%.

When predicting whether lower-court losers like 10X Genomics will lose again at the higher court, Ex Parte has a 93% accuracy rate, says Jonathan Klein, the company’s founder and a former president and chief legal officer at business software company MicroStrategy Inc.

Klein thinks Ex Parte’s data-driven approach will make patent litigation more efficient by giving clients a better understanding of their odds—and which lawyers will improve them most.

“We are able to develop a predictive model that can achieve what even the most experienced lawyer who has argued 20 cases in the Federal Circuit couldn’t possibly hope to achieve,” Klein, 52, said. “That is what’s happening across every other industry but has not yet made its way to the field of law.”

Some lawyers shared deep skepticism about an algorithm’s ability to select lawyers for individual cases or accurately predict winners in high-stakes patent cases.

But attorneys at firms including Fish & Richardson and Sheppard Mullin are using its analysis to win more business or help clients make settle-or-litigate decisions. Litigation funder Parabellum Capital LLC recently entered a “strategic alliance” with the company, looking to make better bets on lawsuit investments.

Ex Parte is in the vanguard of a potentially drastic, if so-far mostly theorized, shift in how lawyers could one day be hired. Hiring decisions in the nearly $100 billion market in which the nation’s 100 largest law firms compete have historically been based on credentials like a Harvard Law School degree or long-term relationships. But many expect that to change as reliable data tracks lawyers’ performance and produces insights about what drives results in more areas of the law.

“Today, clients come to Big Law because we are synonymous with quality,” Oz Benamran, chief knowledge officer of White & Case, wrote in an April LinkedIn post predicting lawyers could one day be hired similar to the way Uber drivers are. “But when our lawyers and matters are ranked for efficiency, the firm’s reputation will become a smaller part of the buyer calculation.”

Years in the Making

Ex Parte’s tool unveiled in February took more than two years to create. Klein, a Harvard Law School graduate, says it is built on a database that includes all patent cases in the Federal Circuit since 2004. There were 4,689 patent cases at the Federal Circuit from fiscal years 2009 through 2018, according to the court’s website.

Ex Parte’s data includes “hundreds” of other variables, Klein said, including a proprietary effort to determine the “novelty” of the patents in the dispute. The patent-specific data includes which patent examiner worked on issuing the patent and who their supervisor was, as well as what area of invention the patent covers. Other companies, PatentVector and Juristat, have created systems that judge the relative importance of patents or track behavioral patterns of individual patent examiners.

Ex Parte’s algorithm also takes into account the Federal Circuit’s 18 judges. It leverages information about their voting patterns, the president who appointed them and where they went to law school. Since litigants don’t know the three-judge panel they will have until the day they show up for oral argument, Ex Parte has a tool to show users their best- and worst-case panels. Klein said he has seen cases where the appellant’s odds of winning according to Ex Parte range from 35% to 58% based on the panel.

Judging Me, Judging You

Goodwin Procter partner Henry Dinger graduated Harvard Law School in 1978 and has been with the same firm for the 41 years since. His firm bio says he is a trial and appellate lawyer who “in recent years” has “concentrated” in appellate matters involving patent law.

But Dinger was not aware when contacted by Bloomberg Law that Ex Parte picked him, out of 24,000 lawyers who have appeared before the Federal Circuit since 2004, as one of the best lawyers 10X Genomics could hire in an upcoming appeal of its $24 million judgment. In front of one hypothetical panel, hiring Dinger would give the appellant a 25% chance of succeeding, according to Ex Parte.

“This is AI gone amok,” Dinger said, noting that he did a “fair amount” of patent appeals work “several years ago,” but is now the firm’s deputy general counsel. “Most of what I do [now] is professional responsibility. Do they have AI for that?”

Klein said it’s not surprising that Ex Parte’s “Recommendation Engine” highlighted Dinger considering his past experience with patent appeals and that his law firm bio still lists him as active in that space.

10X Genomics declined to comment on whether Ex Parte’s odds seemed realistic. A company spokeswoman pointed to a previous statement where the company said it would appeal its loss.

Grading Counsel

Ex Parte’s two-step method to grade lawyers judges them on experience and results. Lawyers can be “class” A through E based on their experience litigating as an appellant or appellee. They then receive a “grade” from 1 to 5 based on their track record. The “best” lawyers are “A5” while the “worst” might be either those that fail to qualify “XX” or have minimal experience and poor results “E1.”

The tool also ranks clients in the same way. Apple Inc., for instance, is ranked “A4” when it is defending a successful lower-court ruling on appeal.

Despite that favorable rating, Ex Parte correctly predicted the iPhone-maker’s district court victory in a case against ZeroClick LLC would not hold up on appeal in 2018. For that case Apple hired Joseph Palmore, the co-chair of Morrison & Foerster’s appellate and Supreme Court Practice and a former clerk for Justice Ruth Bader Ginsburg. Ex Parte gave him a 33% chance of winning. Morrison & Foerster did not respond to a request for comment.

At least three other lawyers would have given Apple a 50% chance of winning, according to Ex Parte’s calculations. Those lawyers were a solo practitioner in Texas named Richard Stanley; Quinn Emanuel partner Robert Feldman; and Theodore Breiner of patent boutique Breiner & Breiner LLC.

None of those lawyers responded to messages about Ex Parte’s conclusions.

Present and Future

Jim Gatto, a Washington-based IP partner at Sheppard Mullin, said he has used Ex Parte to make internal staffing decisions and to pitch clients for work. He called it “unequivocally one of the more impressive products I’ve seen,” and said it fulfills clients’ desires to base decisions on more than gut instincts and heuristics.

“The things that are more repetitive and readily accomplished with technology will increasingly get done by technology, including hiring patent lawyers and firing patent lawyers,” Gatto said.

Still, Ex Parte will have to overcome deep skepticism among some lawyers.

Michelle Armond, a partner at Armond Wilson and former clerk for Judge Richard Linn of the Federal Circuit, said selecting appellate lawyers is a very complex problem in part because lawyers’ track records can be deceiving. Some lawyers try to select cases they know they can win, for instance, while others are taking on tough cases that make new law.

“Artificial intelligence is continuing to improve,” Armond said. “But if you are asking me today, can AI correctly discern the right appellate counsel for every appeal? We are definitely not there yet.”

Ex Parte plans to keep moving forward. Klein said the company has a product roadmap that envisions providing predictions for patent cases in district courts and other areas of law. He wants to be able to tell clients, for instance, what patents in their portfolio provide the best odds of winning in litigation.

Though he says the “Uber for Law” model described by White & Case’s Benamram “won’t happen any time soon,” Klein takes a bullish view on the future of analytics in attorney selection.

“It is a very big market,” Klein said, “and it is just begging for innovation to help it make more efficient decisions.”