Here is the second episode of #TechFoundersTalk, ByteAnt podcast series with successful entrepreneurs and tech founders. Meet our special guest Toby Unwin, a tech entrepreneur, writer, co-founder, and CIO at Premonition.ai.
Yuriy Stakh (Y.S.): - Hi Toby, it’s great to have you here.
Toby Unwin (T.U.): - It’s good to be here. Thanks for inviting me.
Yuriy Stakh (Y.S.): - Let’s talk about Premonition.ai. It is the artificial intelligence system that tracks attorney wins and losses, generating their ratings against judges. How did you come up with this idea?
Toby Unwin (T.U.): - Well, once upon a time, I got sued a lot. I hired dozens of lawyers in my life, I’ve been sued for things like not having an elevator in a one-storey shopping center I owned. Oftentimes, I would hire a lawyer who wasn’t terribly good, even though they were on the Best Lawyer lists, or recommended by a friend, or all those standard ways of choosing. Most of these things have zero correlation with who the best lawyers actually are. Essentially, every lawyer award or rating is ‘pay-to-play’. So it’s a very screwed up system.
Another problem was the data. If I could find case data, then I could crunch it to find out which attorney wins the most and has the best results. It’s a fairly simple idea, but doing it was dramatically harder. There are probably 80 legal analytics companies now, but the difficult part was getting hold of that data. Our two biggest competitors in the market have only 2% of it.
Y.S. - Okay, I see. Actually, you are speaking about your competitors who have 2%, and you have 87%. How come? What is your advantage, is it technology or a mindset?
T.U. - Initially, it was a mindset. Because we didn’t do any product research or market research at all. I just felt that it would be useful, so I did it. And for the first couple of years, there was a massive dose of skepticism. People didn’t believe we had so much data. And only after multiple calls and examples, they thought: Maybe they do actually have what they claim to?
Y.S. - Entering the market is definitely not an easy task. In terms of technology, what was the most difficult part? What did you do to resolve the problem?
T.U. - It was getting the data in the first place. On TV, you imagine a magical database that is up to the sky and has all the information about everybody. But it doesn’t exist. I went to the law library and started looking for cases or attorneys, and I couldn’t find it. So I started looking for the cases that I’ve been in, and I couldn’t find those either. I didn’t believe it. So I searched for a famous case that had to be there, the State of Florida vs Casey Anthony. It wasn’t there. When I asked the librarian about that case, he said: “Not important”. It was on the front page of a newspaper for 3 years, someone died, how is it not important? He said there was no appeal, so no new legal precedents, therefore it wasn’t important.
Therefore, you find the case by going to a specific courthouse. Even though there are 3,124 circuit courts in the US, they are not connected. When I came to the courthouse and asked if I could have the data, they said “No”. Even though these are public records, and they are public servants.
I wound up scraping the data from their website and started adding more courts later. Some of them were really difficult, as the courts become harder to scrape every day. We scrape thousands of courts every hour, in 12 countries. They become harder every day, so it’s a barrier to entry for us.
For example, a few years ago, we got banned by Australia. Within a few hours, we found a way around it, but no one else did. So a few months later, our main competitor started getting that data from us.
Y.S. - Ha-ha, I see. My next question is about people. Above all, you need great people to make the technology work. How long did it take to build the team for Premonition?
T.U. - Getting together all the people took us about two years. We got through quite a few people initially. I got this guy on Upwork, and I had no money at the time. I financed the software by selling one of my watches. It was a pretty rocky start, but he turned out to be really diligent. In 6 years, he’s gone from 3 people, basically working on the sidewalk to a team of 100 people and offices that look like an investment bank. Over half of his people either work for me or someone I’ve introduced. It’s been a really good relationship.
As for the US team, we got them by accident mainly. If you don’t have the time, you just hire the people and see how they do. You hire extra people because half of them won’t work out.
Sometimes, we interviewed the people very well, but the work turned out to be horrible. And then there was another, who interviewed poorly, but aced the technical test and has been amazing.
We also use an amazing system called Teamwork. It’s like a global-shared to-do list. So once people are hired, I give them tasks. And if they do well, I give them more tasks. If not, we give the task to someone else, until they run out of tasks. People are continually being “Soft hired, soft fired”.
Y.S. - As far as I understand, you first started with outsourcing the technical part. What was your experience, did you have any issues with that?
T.U. - Well, practically everything is outsourced. Technically, we don’t have any employees at all. So you just hire contractors and see how they go. Some go well, and some don’t. Once you have the first insights, you can adjust, which is great. We had two teams in India: the guy who was not very professional, and this very professional company who had about 650 employees. They were a bit more expensive, but it gave us some flexibility in terms of cooperation. Every month though they would send ridiculous invoices. Their CFO called to complain about slow payment. I told him we have approximately 80 vendors and they are all paid the same day at the end of the month. Everyone except them. They’d bill us for things they hadn’t done, they’d bill us for things we’d already paid them for. The errors were never in our favor. It took us a month to get it right. In the end, we just stopped using them. Overall, it's very flexible as I can run my business by managing on - TeamWork, assigning tasks, and so on.
Y.S. - Ok, thank you. I would also like to ask about future trends. Let’s talk about AI & Big Data in litigation. How can it or can not change the game?
T.U. - In general, lawyers have resistance to technology. This is why legal tech companies fail. Once, a man came to me with the idea: “I have a system that will schedule depositions and make the document production easy.” And I said: “Noone’s going to buy that.” “Why?” “Because it will save time and money.” Lawyers don’t want to work quickly, that actually costs them money. They want things to go slowly. This is why, oftentimes lawyers don’t want to buy software, and every lawyer thinks they are the best.
However, with the Covid-19 outbreak, everyone starts to move to the cloud. Eventually, people realize that the old ways of doing business don’t work, when you have to be remote. So, there is a drive to adopt more technology and conduct Zoom meetings. In many ways, it’s a lot easier than before. So I believe we are not going to go back to the office, as it’s better & easier this way.
Y.S. - You actually predicted my next question. Do you have any observations or forecasts about how it is going to end up? Will people come back to offices?
T.U. - I think it’s going to be a mix. This is when we get the full advantage of using tech. When we are in the offices, we have this group thinking. Old processes no longer work and we need to learn to communicate digitally. Another important change is that people are making decisions much faster now.
Y.S. - I see. Toby, you founded Premonition in 2014, 6 years ago. Based on your experience as a founder, what advice would you give to the founders? What were your worst startup mistakes to avoid?
T.U. - Well, it’s by doing things and failing them that you learn. One of the problems with MBAs, is they are not taught about failing. The great myth is that if you follow this right kind of business recipe, you are guaranteed to be successful. Except you are not, it’s a recipe that worked for a person in an industry at a particular time. Many of the things that didn’t work for me earlier would probably work now as the market has matured. One of my failures as an entrepreneur is that I’m always ahead of my time, sometimes 20 years or more. Although Premonition is one of my best-timed businesses. If I started it 6 months later or earlier, it wouldn’t have worked.
We tried to approach attorneys early on, but they had no interest in our system, identifying that early was very important. I’m big on dashboards to track & measure everything because if you can’t measure something, you can’t tell what is broken. Admitting when you are wrong is an absolutely useful skill. You can save money and adjust faster this way.
Another issue was with building a crawler for our website. After we started, we realized that it wouldn’t configure it fast enough. So we built a machine that would learn how to crawl a website. We had offline databases, afterwards, we had to move to the cloud, and went through numerous re-writes of our analytics software. Once we realized that we had to rebuild it, it allowed us to go so much quicker. Don’t be afraid to dump stuff just because of legacy costs. If it needs to be re-done from scratch, but it’s clearly the right path, then bite the bullet and do it.
"Don’t be afraid to dump stuff just because of legacy costs. If it needs to be re-done from scratch, but it’s clearly the right path, then bite the bullet and do it."
Y.S. - That's great advice. What tools do you use as a founder that you would recommend to others?
T.U. - I think Amazon Web Services is really good. I like Teamwork, and how it runs. Text expanders are another great tool that I use all the time. It automates my email signatures, greetings, and so much more. In terms of communication, I’m not a big fan of Slack, so we use Skype. Pen and Paper and Goal Sheets are other things I use. I have a 5-year goal sheet. I beat it down into what I might be able to accomplish that month, that week, then that day. I then batch items by their action type, e.g. out of office, web, Word documents, etc, then prioritize the next 2. The first time I wrote out a goal sheet, I put down growing my business (which was in the UK) and moving to the United States. So I noticed that my goals were mutually exclusive, and I had to pick one. I wouldn’t have seen it if I hadn’t written it down.
Y.S. - What was the best idea that helped you to grow your business in terms of marketing & advertising?
T.U. - Well, we had this particular breakthrough. Last year, I was running a lawyer recommendation service for the consumer market. We discovered custom audiences to target. Our advertisement campaign received a 19,8-9% click-through rate when normally you get around 3%. We launched a Google Ads campaign that got us leads for $6 each vs $200 industry average. Now that was a breakthrough. It took us a long time to get the marketing right, but once we did, it brought amazing results.
Y.S. - Wow, that’s amazing. Toby, you are a co-owner, a Chief Innovation Officer of a successful startup. So, I wanted to ask you, where do you take ideas to bring innovation to your business? What inspires and helps you to do that?
T.U. - That is a good question. I like to come up with new things, this is what excites me. Innovation is actually not that complicated. You just take things from another area and see if it works in yours. Or you can combine a couple of things from different areas, and apply it to yours. Another important thing is asking questions that normally people wouldn’t. Like, what would Elon Musk do? What would a caveman do? What would Tony Soprano do? And thus, coming from a completely different perspective enables you to think differently.
"I also hate the phrase “think outside the box”. Because to me, you never have this box. I never have any constraints. So many times, I would come up with the ideas with Big Data and Analytics, because I didn’t know they were impossible."
Y.S. - In terms of technology, which tendencies and trends do you predict to boom in your industry in the nearest future?
T.U. - I wouldn’t say that law is an industry particularly good to watch for tech innovation. There is not much innovation in litigation. Some of the things I do, I take the data and apply it in different fields. For example, the insurance industry is a major user of our data. So they take case types and map them into their business to look from the risk perspective.
And then, I build lots of tools for internal use. For example, we had a CRM that we had to integrate into our system. It wasn’t working for us, so we wound up building our own.
We also use a lot of automation. Once there is repetitive or monotonous work to do, computers are much better doing those things. And a big advantage of robotics and automation is that there will be abundance in things that are going to get very cheap. Like with the TVs.
Y.S. - That’s a great topic to finish with. Toby, thank you for your time and the interview.
T.U. - Thank you for having me. Take care, bye.
and get the latest updates