Starting Anü

An overview of our process and progress in building a professional services marketplace.

Franklin. D. Roosevelt outlined the four freedoms that every human being in the world is entitled to as follows:

  • The freedom of speech and expression.
  • The freedom of every person to worship God in his own way.
  • The freedom from want.
  • The freedom from fear.

FDR had his legacy honored with the 4 Freedoms Park on Roosevelt Island. This tiny floating neighborhood in between Queens and Manhattan is where Max and I met, to begin our entrepreneurial journey, at the Cornell Tech campus. This is where Anü, our startup, came to life.

Entrepreneurship, in my view, is the height and perfect embodiment of these freedoms. As an entrepreneur you build something , in the beginning, that is authentically your own expression. Entrepreneurs undoubtedly exercise more faith than most religious people. If a startup succeeds, an entrepreneur creates wealth for themselves and must always act in spite of fear. This freedom, through entrepreneurship is available to everyone in the world.

In The Beginning

The logo for LexGraft

After graduation Max and I decided to work on our startup LexGraft full time as I discussed here. We chose this name because we were narrowly focused on legal services. Before we got carried away, we wanted to make sure that we were building something people would need. While still in school we were primarily focused on customer discovery — talking to at least 100 potential customers. We were speaking to early stage (pre-seed to seed )tech startup founders about their experiences with finding, using and paying for legal services.

I’ve been in the legal industry for 10 years and my parents are lawyers- so I feel it’s my duty to make a positive contribution to the legal industry.

We found that if a tech startup is part of a reputable incubator or is venture backed, it’s very easy for them to find top tier legal services at a deferred cost. The VC fund or incubator already has a list of preferred service providers for their startups to use. Working with large firms helps the VC’s manage risk from a quality perspective, insurance coverage, geography (the firms have offices in all large jurisdictions) and offering a full scope of services.

However, the startup founders weren’t necessarily happy to go along with these arrangements, but since they were using VC money they had to comply. They didn’t feel they were getting the attention they needed and the legal fees were burning through their cash.

A partner in a large law firm told us that “the startup practice is a loss leader, but we set up the wins for the M&A team when that company exits.”

On the other end, we spoke to small law firms that are building their startup practice. We saw that it is a little more difficult for them to attract many high quality clients because they can’t defer legal fees for a year, with the high probability they would have to write off those fees. In addition, since they specialize in a handful of areas, which normally exclude IP which is important for a tech startup.

The problem we were solving is: connecting bootstrapped startups to small law firms. There is a link between the networks that people are part of and the quality of professional services they get access to. The more elite your network is, the better the quality of professional services you get access to.

This is limiting to both the professionals and the startup because they only get to access people in their immediate networks. It also perpetuates bias, because usually, people are most comfortable with people they share similarities, but miss opportunities to expand socially and professionally.

The Pilot

With this data in hand, we set off on our pilot! Since we were building a marketplace we had the classic egg and chicken problem. We started with lawyers, because I am a lawyer. We called it the Covid Access Initiative where we asked law firms to donate 5 hours to startups for free. We got 4 law firms to sign-up and they were excited to partner with us to discover new clients using our services. With trembling and trepidation we put our services into the Cornell Entrepreneurship Ecosystem to find clients, we had an almost immediate response. It was time to get busy!

The Tech (told by Max)

AI/ML startups be like “Our system is fully automated…”

My thought process for building the technology for Anü started with considering what is important to our professionals when they meet a potential new client for the first time. In our discovery we found that quite simply our partners were tired of using platforms that introduced them to high quantities of unvetted, unserious leads that had to be reached out to individually.

What makes a good introduction and good potential client is someone who is (i) serious about their needs, (ii) is ready to pay, and (iii) easy to get in contact with. These three points were and still are the guiding principles when I think about our technology, and how to build it out.Once we are confident a client passes the first two pillars, we put it in the hands of the client to schedule time with the professional. The third pillar is straight forward. This eliminates the need for the professional to do their own outreach, and lets clients only schedule time when they are sure that they are ready.

How can we use technology to signal whether a potential client is serious about their needs, and serious about their business? Asking explicitly isn’t a solution for the obvious reason that the client can simply lie. That’s why we decided to ask a different question. We asked the potential clients as they came in to give us their best pitch. Tell us about their business, and more specifically what excites them about it, what they know about the space, and what their vision is for it. Our hypothesis is that someone who is excited about their business, committed to seeing it through for the long term, and knowledgeable in their space will make for a high potential introduction, and someone who has a high probability of converting to a long term client. This educated hypothesis was based on discovery with our professionals who consistently described how their favorite introductions were with clients who were passionate about what they do, knew what they were talking about, and hungry to see their business succeed.

The next question we asked ourselves was how can we build a system which evaluates each of these metrics in an equitable, scalable manner. Obviously we couldn’t simply grade each pitch ourselves. Besides not being a scalable solution, our grades would be subject to our own implicit human biases. This made me think about different language models and datasets that I had previously studied, and how we might use these to solve our problem. I landed on sentiment and objectivity datasets, which both perform well under BERT-based NLP models. Given some great open sourced packages which make the training and deployment of these frameworks relatively simple, I was able to build out a model which approximates the metrics that we were looking for.

Upon testing we found that the models performed very well under anecdotal trials with the limited data that we collected. Going forward I’m going to develop these models more as we start to collect our own data, meaning we can rely less on the academic datasets. I plan to write a whole separate post on bias in AI, and how we plan to solve for that. I recognize the importance of not only eliminating human bias, but addressing the bias built into the systems themselves.

The Results:

We were very excited to see that we were solving a problem that mattered. We found that most of our users were female, immigrant and minority founders/teams. We thought that armed with this traction, it would be enough to get some investment, but that’s a post for another day!

Goodbye LexGraft. Hello Anu.

After our investment meetings, we looked at the feedback we received. This year wasn’t kind to the LegalTech industry- with 2 huge blows namely the closure of Atrium and challenges at Upcounsel. Investors looked at us with a raised eyebrow and also felt that the size of the market wasn’t big enough at $1.8 billion. We decided to expand our services beyond discovering legal services to discovering any professional service that a startup would need in their first year of operation — which took our market size to $3.3 billion.

However, we couldn’t do this as LexGraft, so we had to change our name to something new… like Anü.

If you want to hear more about our journey, make sure you are signed up here.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store