top of page
  • twitterX_edited
  • LinkedIn

Jacob Beckerman

Updated: Mar 4, 2023


Jacob Beckerman brings a unique background in finance and computer science to the field of legal tech. During college, he explored natural language processing and video games. After college, while working in finance, he wanted to improve how people worked with and created documents. That desire combined with his expertise in finance and computer science enabled him to develop an innovative tool that seeks to dramatically change how people approach and work with documents.


Tell me a little about your background.


I grew up in Toronto and came to the US to attend the University of Pennsylvania in 2016. I studied finance at The Wharton School (BSc) and Computer Science at Penn Engineering (MSc, BASc) as part of the M&T Program. In college I mostly did my own personal projects and worked in the NLP lab. I always liked tinkering around with AI and video games, and fell into the world of finance through working at Bridgewater the summer between my undergrad and masters.


I worked on CoParse in 2019, spending a year developing the AI component in stealth. In 2020 we raised our first round of capital from BoxGroup, Cooley, Jon Oringer, and others, and got off to the races building the product.


What is CoParse? What is its mission?


We are fundamentally re-building the "PDF reader" and "Word Processor". Adobe Acrobat was released in 1993 and Microsoft Word in 1983. Since then, little has changed. These programs are generalist programs: you can write essays, resumes, blog posts, and more. They are not tailor-made for the complex documents we see in the business world, e.g. financial and legal documents. We are re-building from the ground up using 2020s-era UI, UX, and AI advancements, while tailoring the product for financial and legal documents.


We recognize barriers to change so we:

  • Maintain the .pdf and .docx formats (no new file format or vendor lock in)

  • Have a desktop Windows and Mac application that operates offline (no cloud)

  • Integrate with existing tools and workflows


Our mission, in short:

  • Make delightful software that "just works". Never ship junk. Minimal or no "change management".

  • Price low and reach many people.

  • Deliver returns to stakeholders: customers, employees, investors.


We think of what we're building as an IDE for legal + financial documents. An "IDE" or Integrated Development Environment is what programmers use to write code. It's specialized for the specific coding language and provides debugging, syntax highlighting, multiple tabs, enhanced navigability, and more. CoParse is bringing over many of these features to traditional documents, turning plain text "dumb documents" into interactive, web-like documents.


As someone new to the legal tech world, what are your thoughts on what you have seen and learned so far?


Before coming into this space, I had been told that law was slow to change. On the contrary, I've been super impressed by the people I've met and how they immediately get what and why we're building. Yes, there are barriers, but people by-and-large are excited about change. On the law firm side especially, CIO/CKIOs have very open to our tech and we've had quite a few "intro" calls that turned into pilots or sales, which exceeded my expectations for the pace of change in the industry.


CoParse is also used in banking, investing, insurance, accounting, auditing, executive, and other roles, so I've had the privilege of comparing buying cycles across industries. One thing I like about law is that there's generally a point person in charge of procuring new technology, as opposed to a set of internal buyers. There are still many stakeholders, but the current structure will serve law well as an industry.


The other thing I wondered before CoParse was whether firms that use the billable hour would be open to us, since our 3x-5x efficiency gains would mean less time spent in documents. This turned out not to be an issue.


How did you put together the team you have for CoParse?


Personal networks, mostly. CoParse's new CTO, Sam Kececi, and I worked together at Bridgewater Associates where we pair-programmed (two keyboards, one computer!). It took a while for me to convince him to leave Amazon, where he working in ML Ops, but when he did come onboard it lead to an explosion of feature development. Our other engineers and contractors are friends or friends of friends.


LexFusion has also been a huge addition to the team. People trust them, and I see why. Joe Borstein and I were connected through a mutual friend and he quickly decided it was worth working with us, despite usually only working with mature companies.


Recruiting for any startup is hard, but I think especially hard in LegalTech. The smartest engineers from top schools generally have never heard of the space. Branding matters, and people matter. This is our sustaining competitive advantage. Simply: the best engineers only want to work with the best.


Take Tesla: why is Tesla outperforming in its industry? I'd argue: it's because Tesla is the place that the smartest people want to work. Similarly, we've shown that CoParse is a place that only hires the best, and are turning that into a positive feedback loop.


To someone just starting a legal tech venture, what would be your advice?


To build a venture-backed company: ask yourself, what can you build that can be used by MORE than just lawyers? There are only 1.5M lawyers in the US, which isn't a huge market. So either you need high ACVs (i.e. solve a very big pain point) or you need a bigger market. There are exceptions to this. Think carefully about whether you are building a "product" or "feature".


If you don't want to raise venture money, then the above doesn't hold. There are many "small problems" in the legal space that can be solved by software. They're not big enough problems for venture investors but they're problems nonetheless. Have a network of potential customers who trust you or know someone who does, because trust is key.


219 views

댓글


bottom of page