An Operating System for Startups

A workshop held last night explored how automating legal documentation could be a key technology that helps startups scale across ASEAN. JFDI is spinning out the processes it has developed to address this pain point, writes JFDI co-founder Meng Wong


This post was transcribed from an in-person interview held at BASH at Block 79. Traces of an elliptical, discursive style may remain.

First-time entrepreneurs imagine that cutting their first deal with investors will be a handshake and a glamorous celebration with photographers from Techcrunch queuing up to record the moment. So it comes as something of a shock when they find that reality is 6–20 weeks of tedious phone calls and emails chasing down paperwork, with a mass of messy changes to track on Microsoft Word while the lawyers burn $600 per hour.


During that time, not only does the entrepreneur get completely distracted from running their business but the delay opens up numerous opportunities for the deal to fall apart. An investor can turn around and say: “So sorry, but I just got hit with a tax demand … or an invoice for school fees … or my daughter is getting married and I want to use the cash for something else”. Or SARS happens, or a financial crisis in Greece, or the Chinese stock market drops 30%. Suddenly the deal is gone.

It’s just as bad for investors, in a different way. Instead of spending their precious time finding great deals or adding value to the teams they in their portfolio, they have to wade through miles of fine print looking for errors and content left over from the last time they used the template. If a friend wants to come in on the deal, they have to go through it all over again.

Multinational Muddles

Now imagine a large portfolio with investees scattered across ASEAN where every country has its own legal jurisdiction. At JFDI we have one of the largest seed-stage portfolios in Southeast Asia, having funded more than 60 startups. We have a duty to track every single one of them on behalf of our shareholders. That’s hundreds and hundreds of contracts. It only gets more complicated as the startups start scaling out across the region. Instead of enabling companies to scale internationally, the well-intentioned detail of legal documentation ends up piling on cost and causing delay.

JFDI has developed a lot of legal infrastructure internally to deal with this. As more accelerators and crowdfunders have set up nearby we can see that everyone is facing the same challenge, or soon will. So, in the spirit of sharing, we are exploring ways to help everyone avoid reinventing the wheel. That’s why I spoke yesterday to our fellow investors at BANSEA and TiE Singapore about the work we’ve put online.

A shared pain point

Angel investors face a learning curve just as steep as beginner entrepreneurs. When you’re investing $100,000 in a startup, you don’t want to spend $40,000 on legal fees. But you can’t afford mistakes, either. Early stage investors must go for volume to succeed.

To ease this burden, everyone from Y-Combinator, Techstars, 500 Startups and the US and British National VC associations have published their own template termsheets and investment agreements. In fact there’s been an astonishing proliferation of these things. Today dozens of websites offer legal templates: Docstoc, Legalzoom, Rocketlawyer, DragonLaw, Lawcanvas just to name a few.

Adding Functionality

So this project is a natural evolution. With a group of volunteers we have imported every freely-available contract template that we can find into a rudimentary template filling engine. So you can enter your information once and then press a button and compile that into whatever agreement you feel is most appropriate. This is especially useful for situations like fundraising where dozens of documents may be involved in a single transaction.

Our prototype is called Legalese and currently covers incorporation and fundraising for angel and seed rounds, plus volunteer agreements, internship agreements, and directors’ resolutions. Over time we plan to add HR contracts and everything else an early stage company needs.

A vision for Asia

Legalese has received a particularly warm reception in Asia because most of the template contracts are written for US jurisdictions like Delaware. That’s fine if you’re an American lawyer; as someone observed, the size of the litigation market in the US alone exceeds the GDP of Southeast Asia.

But the Internet is multinational and multilingual. And so is the startup world. We would like to build an ‘operating system for startups’ that helps anyone, anywhere in the world, to incorporate digitally and then strike contracts as easily as connecting through an API online. We plan to launch with Singapore law first, then US law and then scale out to include every other jurisdication in the world.

There are two ways to do this: spend millions of dollars on hundreds of lawyers; or spend nothing at all, relying instead on tens of thousands of opensource contributors.

Insights and Innovation

Lawyers are like programmers – except they’ve been trying to program, for the entire lifetime of their profession, without a programming language. In the next few decades, that will change. Software will eat law. Coders will start producing contracts without ever involving a lawyer.

It’s obvious to any computer scientist that we should be able to write contracts in programming language. This impulse comes very naturally to hackers and coders – the notion of writing a contract in Javascript and then compiling it into whatever jurisdiction you want, just as websites automatically serve content in multiple language. We’re working on that: it’s a research project within Legalese.

Another implication: social practices around software – how end-users discover, share, and support software; how developers test and contribute improvements; how businesses monetize – all of these practices are going to change the future of law. I’m not alone in saying this. Yesterday I attended the Singapore Legal Futures Conference where thought leaders like Richard Susskind, author of Tomorrow’s Lawyers, were singing the same tune.

No more hand coding

Right now you pay a lawyer $600 an hour to check clause numbering in Microsoft Word. Yet there are still mistakes because they are doing it by hand. They don’t see any need to change because they are paid very well but enlightened consumers of legal services are beginning to insist that there has to be a better way.

We want to make it really easy to do things that you could not imagine a traditional lawyer doing. For example, a founder might say “Right I have a YC series AA agreement that I would like to shop around to investors”. If the investors come back: “We like the general idea but would prefer to do this using a KISS investment structure”, (by 500 Startups), that’s no problem. It will take 5 minutes to turn that around online, whereas a traditional law firm might take 2 weeks and $6,000.

And because this is open source, if someone finds a mistake, then it gets corrected, and the correction benefits everyone who ever uses that agreement again. We take that kind of thing for granted in software development but today’s law feels like we are still hand coding assembly language. It’s time to leave that behind and let the machines do it right, consistently, every time, faster and at dramatically lower cost.

Coding for the first computers required programmers to rewrite from scratch for every manufacturer’s machine. Today, we have operating systems that abstract away the drudge work of conjuring up windows and handline hard disk operations. Compilers make it easy to port applications from one machine to another. Yet when we move a business from one country to another we pay lawyers to rewrite all the documentation manually. Why?

But is there a business here?

Having screened thousands of applicants to JFDI, I’m become acutely conscious of the difference between a science project and a startup. Right now Legalese is a science project, in customer discovery phase, working toward problem/solution fit. If people are interested in using it on commercial terms then we will incorporate it, raise funds, start monetizing. But right now, we have no revenue, no funding, no salary; we’re just an opensource project, one among millions. How can we afford to proceed at such a measured pace? Because, just as there’s only one Wikipedia, and only one Github, there will be only one Legalese.

Write a Comment

Your email address will not be published. Required fields are marked *