Why Intraspexion’s Use of Deep Learning May Be Deeply Valuable


Editors note: This past month, AI Trends editors spoke with Nick Brestoff, the founder and CEO of startup Intraspexion. That interview was interestingly enough the “most viewed” article of the month even though it was published on August 22. In the interview, Brestoff said that his software system, powered by a deep learning algorithm, could help companies avoid litigation, and save millions. Since we did cover machine learning in the legal industry a few times, and each article generated significant interest, we decide to go back to Nick and learn more.

AI Trends: Nick, what is the exact business case of your platform?

Brestoff: I started with reputable cost data from Towers Watson, which is listed on New York Stock Exchange as TW. For 15 years, TW had tracked the litigation cost for commercial tort litigation, meaning litigation against businesses, and personal tort litigation, meaning automobile accidents.

AI Trends: For which years?

Brestoff: TW stopped after its 2011 Update, and I used the last 10 years, from 2001 through 2010. I used only TW’s costs for commercial litigation. By “costs,” TW meant payouts for settlements and verdicts, defense attorney fees, and administrative costs.

Brestoff noted TW obtained its cost data from A.M. Best and SNL Financial. But that’s as far as TW went, so he went after the caseload for that same 10-year period.

AI Trends:  Was there a particular motivation?

Brestoff: In part by a book contract.  When I was writing Preventing Litigation: An Early Warning Signal, Etc., which Business Expert Press published about a year ago, I knew the business community wouldn’t listen unless I took the next step. So I graphed TW’s data and came up with an average cost of $160 billion per year, which, for 10 years, is the staggering amount of $1.6 trillion. Then I researched the federal and state caseload against businesses. My sources were the judiciary’s federal litigation database called the Public Access to Court Electronic Records or PACER, and the National Center for State Courts. I came up with a little over 3.9 million lawsuits for that same 10-year period. And then it was just math. I had covered the topic in the first nine chapters of Preventing Litigation. But it was still a scary calculation.

AI Trends: That sounds like a pretty high average litigation cost?

Brestoff: Since I was dividing $1.6 trillion by 3.909 million lawsuits and came up with an average of just over $400,000 per case. Actually it was $408,000 per case. But because the federal and state court databases are not in sync with each other, and I had to make estimates and wanted to be conservative, and so I backed off by 15% and rounded off to $350,000 per case.”

AI Trends: When we spoke last,  you claimed that Instraspexion can save millions in litigation costs?

Brestoff: We calculate that by avoiding only three average lawsuits, a business can preserve over $1 million

AI Trends: Give us some insight into large company potential risks associated with internal communications?

Brestoff: IntraSpexion solves an enormous enterprise pain. In-house counsel is blind to internal communications that are risky. Because we learned how to use deep learning, the blind can see. Deep learning in this application for business is a deep disruption, a bottoms-up disruption. We won’t be in a price war with other legal tech offerings, most of which are focused on eDiscovery and so are after a lawsuit arrives on a company’s doorstep, and we aren’t offering a more efficient way to track or understand the language of contracts, legalese, or a better way to do legal research.

AI Trends: Thank you Nick. We’ll let you know how many views we receive with this interview!

Editors note: Intraspexion will be among the exhibitors in the Emerging AI Products Pavilion at AI World, and will participate at the AI Technology Solutions Theater on November 8.  For more info, see www.aiworld.com