What if lawyers could read a judge’s mind?
A startup says its AI can get close, promising lawyers a glimpse into how judges think before a single argument is made. That’s what investors are buying into with legal tech startup Bench IQ, which just raised $5 million to bring the idea to market.
Jimoh Ovbiagele, the company’s cofounder and chief executive, said Bench IQ has built a proprietary dataset and layered in large language models to forecast how judges tend to think and rule. He wouldn’t disclose the sources, citing competitive reasons, but the company’s terms of service suggest the data includes federal court transcripts.
Attorneys start with a simple query: How has this judge handled emergency motions in the past? Behind the scenes, Bench IQ says it spins up thousands of virtual assistants, or “agents,” to comb its data on the judge. Within minutes, the lawyer gets a report that answers the question, citing the judge’s past rulings.
The tool aims to become a cheat sheet for the bench, and the team behind Bench IQ, which includes Jeffrey Gettleman, who spent 17 years as an attorney at powerhouse law firm Kirkland & Ellis, thinks every litigator will want one.
Ovbiagele said users include four of the top five Am Law 200 firms, the nation’s second-hundred largest law firms by revenue. Cooley, Fenwick & West, and Wilson Sonsini, three premier law firms for tech companies, are also early investors.
The idea of selling lawyers a crystal ball isn’t new. LexisNexis and Thomson Reuters, the owner of the legal research platform Westlaw, already offer tools that chart ruling patterns, favored precedents, and timelines for cases before a given judge. Theo AI, a newer entrant, pitches its software as a way to predict settlement outcomes for Big Law firms and corporate legal departments, building custom engines trained on case histories and proprietary data.
Ovbiagele argues that incumbents can only show how often judges rule a certain way. Bench IQ seeks to uncover why judges make decisions and how lawyers can influence them. To explain the difference, Ovbiagele offered an analogy.
“Knowing how often a judge rules a certain way is knowing how often wildfires occur in a region,” he said. “It is useful, but it’s limited. Using Bench IQ is understanding why those wildfires happen and how to avoid them.”
Investors are betting there’s a big market for that kind of intelligence. Bench IQ told Business Insider it has closed a $5.3 million seed round co-led by Battery Ventures and Canadian venture capital firm Inovia. Existing backers Haystack, led by Semil Shah, and Maple, led by Andre Charoo, also increased their stakes after leading the company’s $2.1 million pre-seed in early 2024.
Today, Bench IQ focuses on federal courts. Ovbiagele says it wants to use the fresh funding to expand its dataset to cover state courts and hire more engineers and customer support reps.
‘Unfinished business’
Bench IQ is entering a crowded field of startups chasing law firm budgets. It will have to vie for customers against platforms like Harvey and Hebbia, which promise to strip away the drudgery of legal work, and startups targeting perennial pain points like timekeeping and billing.
Ovbiagele has familiarity going for him — or notoriety, depending on who you ask. Thomson Reuters sued his last company, Ross Intelligence, into oblivion.
A decade ago, Ross tried to speed up legal research with artificial intelligence. In 2020, Thomson Reuters sued the company, accusing it of using Westlaw’s content to train a rival tool without permission. A federal judge later ruled that Ross’s use of Westlaw headnotes wasn’t “fair use” under copyright law.
Ross shut down soon after; Ovbiagele says the litigation “bled” it dry and made it impossible to attract new investors. The company has since appealed the ruling.
A spokeswoman for Thomson Reuters declined to comment.
The shutdown left Ovbiagele with “a burning sense of unfinished business.” With Bench IQ, he has another chance.
Have a tip? Contact this reporter via email at mrussell@businessinsider.com or Signal at @MeliaRussell.01. Use a personal email address and a non-work device; here’s our guide to sharing information securely.
