A couple of days ago I saw this @ronfriedmann tweet:
Interesting R&D + practical transaction resource
— ronfriedmann (@ronfriedmann) June 30, 2015
So LexPredict enters the we’ve-got-lots-of-contracts space. Ron will of course guess my reaction, which is … lack of excitement.
Yes, there are loads of contracts out there. Yes, we can data-mine them, and it can be useful to do so. But the challenge facing anyone picking over contracts in the great boneyard that is EDGAR and any other comparable database is how to separate that which is relevant and high-quality from that which is irrelevant and crappy. It’s not a task for dilettantes: the more you need stuff from EDGAR, the less likely it is that you have the skill necessary to look out for yourself.
So that’s why I’m not excited by LexPredict, just as I wasn’t excited by Bloomberg Law’s Corporate Transactions (see this post).
But instead of continuing my spoilsport routine, I’ll now hand off to Preston Clark, he of LawInsider.com, which I wrote about most recently in this December 2014 post. Because LawInsider.com also mines EDGAR, I asked Preston for his thoughts on LexPredict, and here’s what he said:
I’ve not spoken with Michael Bommarito or Daniel Martin Katz, but obviously the founders have a strong track record in legal data technology and the legal tech space. Similar to what we’ve been working on over at LawInsider.com, LexPredict appears to be venturing down the path of creating machine learning algorithms for the identification and extraction of contract metadata. Based on their reported contract sample size (400K) and the date range (20 years), my assumption is that they are pulling from EDGAR. The SEC data set is literally millions of pages of contract language, and this presents a tremendous machine learning opportunity. It’s a great project and I’ll look forward to trading notes with Michael and Daniel at some point. The bigger question will be their monetization model, which, based on the website (“Customers can choose between one-off historical samples or ongoing updates.”), seems similar to what TechAgreements has been doing for a number of years, and which I perceive to be a difficult path. Hopefully I can talk them out of that when we speak.