Robot picked its name and bought LegalRobot.com — How Legal

An end user explains how they picked their name and domain name.

This is the first of what I hope to be many profiles of firms that have purchased domain names on the aftermarket. This series will help people understand the naming process firms go through, how firms view domain names and domain sellers, and what the purchase process is like in the buyer’s view.

Today we’ll look at Legal Robot, a fresh service from Agnito Labs, which purchased LegalRobot.com for $1,200.

DNW: What is Legal Robot?

Dan Rubins, CEO & Founder: Legal Robot is Artificial intelligence for legal documents that helps people understand legal language by providing an instant breakdown and error check for contracts, helps people write better legal language, and lets businesses close deals faster.

So much information is locked away in convoluted legal language, accessible only to those with special training or those with the means to hire them. Legal Robot uses machine learning to instantly and inexpensively unlock this complexity for consumers and businesses to improve both efficiency and access to justice. We are building the next generation of Legal Technology that will improve trust and fairness, as well as transform the way people use the law to interact with each other.

DNW: What was your name selection process for the service? Did you consider lots of domains? How did you ultimately decide on Legal Robot and LegalRobot.com? Did you consider alternatives to a domain ending in .com?

Rubins: We started out with Contract.AI – the domain was available and perfectly described the product at the time – Contracts + Artificial Intelligence (AI). Since the word “contract” is so general, the domain extension was really the distinguishing part of the brand. Ahead of long, our product idea expanded and we realized the product shouldn’t be limited to just contracts and didn’t want the name to hold us back – so we started looking again. We actually considered at a lot of AI domain names thanks to their availability and fit with our concept but also broadened the search to .com’s. We started collecting ideas and fitting concepts together into a big list of names (aided by namemesh.com), then started crossing off names belonging to actual firms, unresponsive sellers, and squatters with unrealistic expectations. We finally narrowed it down to just a few names based on how the domain name fit with our concept. LegalRobot.com was our favorite and the seller was responsive. A couple months later, we heard regarding one of our few (and more innovative) competitors – beagle.ai – who decided to stick with the .ai extension.

As a tech startup, we can get away with a lot of domain name shenanigans – .io, .ly, .co, .ai … but .com is still the best choice (if you can get it) since it’s the default in most people’s mind. I still think .ai domains are still a great option for Artificial Intelligence startups – especially those looking for generic names or those with name conflicts. Still, there is so little competition from .ai registrars that it can be overly difficult and expensive to get a .ai domain name. Perhaps we’ve been spoiled by the thousands of registrars offering common TLDs that we forget how this industry can be so archaic at times.

DNW: Since the domain name was by now taken, you had to buy it. What was the purchase process like? Do you think you paid a fair price?

Rubins: The purchase process was fairly simple, the domain was listed with sedo.com, I submitted a pretty low offer and got a counter offer of $1200. The price seems quite fair in retrospect now that it has become an notable part of the brand, but at the time it felt like highway robbery. Sedo handled escrow and payment verification; the process was wrapped up in a few days.

The whole pretense of having a domain listed without pricing seems weird to me – are you selling it or not? If the seller’s goal is to complete the transaction, why put up an artificial barrier; is the offer/response smokescreen really going to improve pricing so much that it outweighs conversion?