Andrew Einhorn Podcast Transcript
Andrew Einhorn joins host Brian Thomas on The Digital Executive Podcast.
Welcome to Coruzant Technologies, Home of The Digital Executive Podcast.
Brian Thomas: Welcome to The Digital Executive. Today’s guest is Andrew Einhorn. Andrew Einhorn is the Chief Executive Officer and Co-Founder of LevelFields, an AI driven fintech application that automates arduous investment research so investors can find opportunities faster and easier.
His mission is to create AI tools that make advanced financial strategies effortless and accessible for all. With over 10 years of experience in building and leading technology and tech enabled service firms, Andrew has a proven track record of delivering innovative solutions that solve real world problems and generate value for customers and stakeholders.
Well, good afternoon, Andrew. Welcome to the show!
Andrew Einhorn: Hi, Brian. Thanks for having me!
Brian Thomas: Absolutely. Appreciate you making the time, Andrew, today. This is awesome. Love starting the day out with a podcast and jumping right into your first question. Andrew, what inspired you to start Level Fields and how did your background in technology and tech enabled services influence your vision for the company?
There’s a long story in the short version I’ll start with the short version of that one. I had just sold my company, my previous company, which also did kind of event monitoring analysis for publicly traded companies in 2019, took a bit of a break, went on what my wife called a sabbatical. And then about six months later, kind of regrouped with the technical team.
And we were thinking about how we could improve 10 years, take it to the next level. We had seen kind of the limitations of Boolean searches and knew we needed to add an AI component. So, we started building out an AI system that was much more advanced for linguistics processing. And as we were working on that, the pandemic hit, COVID happened, and the market started selling off like crazy.
I started getting lots of calls from family members who were concerned about, you know, losing their retirements forever, losing their college savings for their children. And there was this just deep fear in the marketplace that we were headed into a global depression. My background was in epidemiology before I got into technology development, and I worked in public health for 10 years.
So, I had seen pandemics before I’d actually worked at the CDC. As a contractor during the swine flu endemic, I just decided to start doing some research and looked at all the past pandemics and how they affected markets looked at similarities that might exist between kind of the Zika epidemic that we saw.
And others looked at the government responses and locked myself in my office for about 80 hours to come up with the answer, which was markets will be right back to where they were in 6 months. And the government response will be swift, and we’ll have a vaccine within 13 months. That was what I told people.
I thought I was crazy when I said it and I dismissed it, but, you know, lo and behold, that’s exactly what happened. And it wasn’t terribly surprising because that’s what the data showed. And so all of that kind of showcased that events change everything. They change the trajectory of the stock market.
They change the trajectory of a company’s revenue or history. And once we had realized that we saw, you know, what, There is an opportunity here to make this kind of information available at the touch of a button, you know, it shouldn’t take 80 hours to do this kind of work. And the more we thought about it and kind of peeled back the onion, we saw, wow, there’s so many events in the market.
And then surrounding these events are just a lot of opinions. How can we, we through this noise, you know, all these opinions and all these kind of influencers and pump and dumps that are happening in the marketplace and really just focus on the core of what’s moving. Financial markets and stocks, which was events and so that was kind of the aha moment.
And, you know, then we just started thinking about what the user interface should look like, how we make this as easy as possible. So that it’s like, looking at a weather forecast, you know, you can see. Hey, you know, it looks like a hurricane is going to come, but it’s not going to hit till Thursday. And when it does, it’s a category 3, and this is how much damage it’s going to be.
We wanted to do the same thing with events, right? We wanted it to be just that easy, that straightforward. And the more we kind of leaned into that, the more we recognized there were great trading strategies and investing strategies that could be based solely off events. And you could look at one type of event that occurs again and again and again, you know, a couple hundred times a year and just invest or trade based on that event, you wouldn’t have to know nearly as much on fundamental analysis or technical analysis.
So, the kind of barrier to entry was lower and we felt like, Hey, this is a really opportunity to kind of take what the largest, most profitable, most successful. Hedge funds do by leveraging events and democratize access to that so that any person can have the same level of insights and technology without having to work on wall street for 20 years.
So, that’s where Level Fields was born.
Brian Thomas: Thank you, Andrew. That’s a great story. And obviously we could learn a lot from the pandemic, especially since a lot of research has come out after the fact now. And, you know, people are really looking at really what happened, but I love the fact that you jumped in there and is leveraging AI and predictive modeling to forecast events and what’s may happen in the market. So, I appreciate that.
Andrew, LevelFields leverages AI to automate investment research. Can you explain how your platform works and the specific problems it solves for investors?
Andrew Einhorn: Sure. The short version is it’s like a speed reader. So, think of it like an AI is, is a robot that reads really fast. So fast that it can go through 30, 000 documents every minute.
And it’s looking through, you know, millions of records that are out there in the stock market. News, company announcements, filings, things of that nature. Some of which are really detailed reports, some of which are one page long. And all of those, they hold different pieces of information. We specifically look for events, things that happen.
You know, an announcement might be a product launch. Could be a company that’s returning capital to shareholders. Could be a CEO getting fired or getting hired. You know, we look for those kinds of events that are material to the share price, meaning they will move the stock price up or down and we extract them.
And so, you know, if you ask a typical person how many stocks can you name, they’ll probably come up with 20, maybe 30. You get someone more experienced; they might be able to do 50. A few people will be able to list the 6,000 stocks that exist in the U. S. stock market alone. And so that is the problem that we’re solving for is that a single human cannot monitor the market without the help of, you know, hundreds of analysts like they have at big asset management shops.
With our AI, you can actually do that. You can monitor and read through all these reports looking for just to serve information that you’re interested in. So, I can set an alert with level fields. A. I. to go out and say, you know, I’m, I’m really interested in tech companies. That are medium sized that have a P ratio below 20.
That just got a new CEO, because I think that those are good turnaround stories. And it just looks for that 24 hours a day, seven days a week. Once it finds that particular match of the criteria, it will flag it, send you an alert and give you information about the company. And so your research time goes from, you know, 10, 20 hours a week to about 10 to 20 seconds, and you can do that for any variety of different types of events.
Or you can just log on to the system and search and browse through an array of events that are happening to kind of get a sense of what’s happening across the whole market. Just as easily as you can go browse flights and find a flight that’s going to a place that you’re interested in. So, we really try to think about the end user and the problem that most people have is that they don’t either have the time to do the research in the market.
They don’t have the knowledge to do research in the market. You know, or they simply just cannot keep up with the volume of information that’s out there. So, we aim to leverage the AI to solve those problems and make it really less burdensome to come up with different investing ideas, to come up with trade ideas, to keep tabs on, you know, what are the kind of emerging companies.
You know, hidden nuggets in the market that are having an odd, you know, tailwind due to some macroeconomic factor, you know, for example when the war in Ukraine broke out and Russia invaded Ukraine, we were getting these event alerts for lots of fertilizer and coal companies that were handing back tons of cash to shareholders.
They were giving and raising dividends, they were doing stock buybacks, and it was just sort of odd when the rest of the market was selling off because 2022 was pretty bad. And so, it forced us to sort of look and squint and say, what is going on with coal? What is going on with fertilizer? No idea. Well, because Russia had cut off natural gas.
Most of Europe had to switch to burning coal to produce power, which created more demand on coal, which drove the price of coal up. So, if you’re a coal company, and you can now do nothing different, but sell your, your product for 3 times the price. You’re going to have a lot more money, and it was the same situation with fertilizer.
A lot of it came out of Belarus and Ukraine and Russia. And because, you know, the Russian ships were, you know, Basically, in in the ports, you couldn’t get this material out. And so, companies who were making and distributing fertilizer from Canada and the US could just sell their product for a lot more.
And I would have never thought of that. In a million years gone and actively search for fertilizer companies. But because I saw these special events that were happening. I had to pay attention to it ended up investing in in some of these. And made a good amount of money, you know, just by having the system kind of find these anomalies for me.
So, you know, those things are happening all the time in the market. We always have a lot of. Unique events, particularly, you know, in the last 4 or 5 years, we have constant macroeconomic events. We have, you know, political turmoil and we have global conflicts and we have weather issues. There’s no shortage of reasons that the markets are reacting and.
The difficult thing, I think, for most people is to kind of put 2 and 2 together across. All of these companies and understand the economy. You know, at large, when there’s 6, 000 different participants in it, and, you know, what’s going to cause the price to go up and down. And that’s, that’s what the system is doing automatically.
So, we’re trying to really break down the barriers, make it easier, make it more visually appealing, take kind of an Apple approach to, to design. And we’ve seen, you know, really good support for, for what we’re doing. We’re not perfect, but, you know, we keep, you We keep improving every week and we’re about to launch a whole new slew of really cool features. So, I’m excited about it.
Brian Thomas: That’s awesome. Thank you for that, Andrew. Obviously, the world is a volatile place. The last three to four years have been kind of a roller coaster ride, to be honest. So, I appreciate you sharing that and sharing the fact that your technology can find those anomalies and help people make better investment decisions.
So, Andrew, you designed the first AI event intelligence platform to forecast the impact of events on stocks. Can you discuss the development process and any key challenges you face when developing the platform?
Andrew Einhorn: Sure. Yeah, happy too. So, the First thing you have to start with is focusing on the problem at hand and not trying to do something that’s too broad. A lot of off the shelf AI systems. Are not particularly intelligent, right? They’re kind of the starting point. They’re a sketch when you need a painting. And so, you still have to do a lot of work. If you’re trying to use something like that, because it’s, it’s just general information. So, what did we have to do.
We started from scratch and build our own library of terminology of phrases that is specific to the finance sector because they use different language. Even just understanding the way that they use language earnings. Adjusted earnings, diluted earnings core, you know, the things. That are unique to the sector had to build out a pretty massive ontology of terms and phrases.
And that was kind of the starting point and then you have to look at. Well, what is going to trip the system up, right? Because computers and AI, they’re not particularly smart at the beginning. You know, they’re smart after many, many, many, many times training it. And I often say, you know, if you’ve ever trained a dog, it’s very similar to training AI in that, you know, you can train your dog to not be in your house.
But the second you take your dog to your friend’s house, the first thing it does is pee in their house. And it’s because the dog only learned over and over and over again to pee in your house. If you have never taught it, don’t pee in any house. Training AI is the exact same way. It has to, you have to train it for every possible situation.
You have to train it hundreds of times for it to learn. You can kind of skip some of that by having more focused rules and more focused you know, financial language. So that’s a challenge in and of itself just to get enough volume of data out there to process to look at. To train, which takes time.
And then then you have to look at the things that trip up the system where you have a blue bird that flies in the sky. And then you have a blue bird bio, which is a publicly traded stock. So how does the system differentiate between the 2? You have to use context. You have to develop different algorithms for probability assessment of understanding, you know, who is the speaker?
What is the situation we’re talking about? Are we talking about nature? Are we talking about finance and biology? And so you have to bring that context to life. You have to look at things like directionality, you know, so if you’re talking about 2 companies and one is buying a company. Who is buying who and that can get, you know, a little dicey because the company who’s being bought might say, hey, we were purchased by Apple and Apple might say, you know, we bought this other company.
So, you have to understand kind of who is doing what is doing the action and just it’s a variety of rules like that in linguistics. They have to go through apply. And then you watch to see how the system handles it, and you train it and you correct and you continually improve that cycle over and over and over again tens of thousands of times.
And so, we did that, you know, in the course of a couple of years. And so that is on the side of it, but then once you have that, that’s great. But then you have. You know, how is a human going to interact with the system? Right. I’m a user. We launched a beta version that we thought was, was really good. And it turns out from, you know, the feedback that people just didn’t understand what to do with the system. Wasn’t obvious to them. They wanted to go and look for a stock that they knew. So, they wanted to search that’s, that’s how they wanted to start the process. And we didn’t want to look at it that way. We’re like, why would you want to search for a stock? And we can give you all these other ones. That is better that you’ve never heard of, but what we found was people had behaviors and they did not want to change those behaviors.
It didn’t matter that we were coming up with something completely new or a new way of doing something. They needed to do what they normally do, and just have a tool that kind of caters to their process rather than jump in and take our process. And that was that was a big lesson learned on this is that, you know, we, we couldn’t completely change the process. We couldn’t kind of reprogram the way that people behaved. So, we had to sort of find different ways to introduce our concepts to a normal workflow that they were going through. And so, we introduced the search feature and then people would search for the stocks that they want, and they’d find the stock that they want, and they could see.
The events that we would show that were happening to that company, and then they learned, oh, I’ll get it. Okay. This is different from just looking at, you know, a news feed. This is an event. And this event has a forecast of, you know, how these events usually play out and it has you know, an event impact that’s really interesting and they, that allowed them to then understand the platform a bit better and begin to form kind of a rationale or a process for using it.
So that, you know, that discovery process that just went through and the user interface, you know, that was like a 6-month time of learning. Reworking it, going back to the drawing board. Adding quickly some of the features that people expected, testing it, seeing if you know, it caused less anxiety for people and they could adopt it faster, which they did. And then we had to build it out and launch the next version.
So it was, it was interesting. And I have to say that within finance and like some of the other areas that I’ve worked in people are often set in their ways and, you know, they’re not. Very few people look at investing or trading the same way, like, we’ve had thousands of users subscribe to the system and we’ve talked to hundreds of them.
I don’t think any two people had the exact same process. I mean, you know, one person says, “Oh, I use options and then they use calls.” Another person sells puts another person buys puts and another person does strangles and straddles, and another person only wants to make one percent a day. So, they come in and out of the market.
Then someone else wants to automate it. I mean, it’s just an endless array of this is how I do it. And this is the right way. So that was a challenge too, because we had to keep this, this platform so flexible that everybody could get what they wanted out of it without overloading it with features and making it feel intimidating.
And I’d like to say that we completely solved that problem, but I don’t think we have completely solved it. I think that we’re still trying to find different ways to make the platform, let’s say, certain, certain features appear for different levels of users and working on that challenge. You know, all the time and love to get feedback from users who can say, “Hey, I don’t know how to use these filters. So maybe you could just give me like one easy button that I can push, and the magic happens”. So, things like that I think are a little bit unique to this. I think the last barrier I would say is that in this space, unfortunately, there have been a lot of scams.
There have been a lot of false promises, false prophecies that, you know, this company can figure out who the next Amazon is, and nobody can do that, but people fall for it. They put their money in, they lose their money. And so, it’s been a little bit of a challenge on the marketing side to say, Hey, we, we understand that there’s been kind of scammers out there and lots of opinions out there that that say that they could help you become the next millionaire.
We actually have a system that works, and the data is shown right there on the system. And that’s been a bit of a barrier, you know, to, for the selling process to say, look, we, we have something really interesting here. That works. And there’s a lot of like skepticism around that. You know, we do what we can on the lots of case studies and demos that we put on the website, but That’s still going to be, you know, an issue, I think, unfortunately, as people have a tendency to take advantage of other people as sad as that is, it’s just kind of the world we live in.
Brian Thomas: Thank you. Andrew. You’ve covered quite a bit there. Of course, the challenge you’ve had was creating the platform and how AI learns right through that, that long LLM process and need specifics to do things right. But like I said, covered quite a bit of ground there.
Switching to the last question, Andrew, if you could briefly share looking ahead, how do you envision the future of FinTech and the role of AI in transforming investment strategies and financial services?
Andrew Einhorn: I think it’s going to be kind of bifurcated between enabling a person to do a lot much faster the way that LevelFields works where, you know, one person can really now do the work of 100 people. I think that’s going to be a big piece of it. And so, you can take a system and just really streamline your research process, streamline your discovery process.
I think that that’s going to be, you know, an interface like ChatGPT, I’m sure will be layered over financial data. So, you can just kind of ask questions that are more specific and a little bit timelier, like, you know, Hey, why was Apple up 7 percent yesterday and get an answer very quickly, not have to poke around and I think that will help.
On the other side of it, you know, and this has been going on for a while and the asset managers are continuing to look for ways. To automate the entire process of portfolio management, wealth management, and to look to. AI as a means to sort of automatically. Balance portfolios, right? So that means if.
If they’re supposed to be 10% of the portfolio tech and the tech companies that they’re invested in become 20%, then the AI will automatically sell off some of those assets to get the balance back to 10%. And there’s some of that that, that has existed for a while. So, I think it’ll just get more advanced.
And some of the things. It’s looking for, and I think some of the top hedge funds are certainly going to try to find different ways to look at, you know, using some of the momentum in the market to identify better stock picks over the long term, based upon early indicators that AI can analyze in a multivariate fashion, you know, looking at multiple variables simultaneously to try to project, you know, where companies financials might be in a few years or where a country’s growth areas might be based upon looking at, you know, other countries that had similar trajectories. So, it’s not dissimilar from what we’re already doing. It’s just more efficient. And the analogy I like to give is, you know, once upon a time, if you’re going to build a house, you had a hammer and some nails.
Now, you have a nail gun and AI is pretty much a nail gun. For any kind of thing that you’re trying to accomplish, it doesn’t mean that you’re going to eliminate people altogether. Although in some industries that might happen for the most part, it’s just more efficient. You can do a lot more faster. And I think for a country that is facing population growth problems, like we are here, that’s good news. It means we can continually enhance our productivity without enhancing our population size.
So, and I’m excited to see what comes next. You know, we’re certainly trying to participate in this by. Having homegrown AI, other fields, and hopefully people find it as useful as we do.
Brian Thomas: Thank you, Andrew. I appreciate that. And I always like to get the guest’s perspective on the technology platform or, or whatever, you know, technology they’re actually working on or developing. And so, I love your perspective on that. And I think AI has a lot of promise. We just need to make sure that we keep the ethics in, in check. Andrew, it was certainly a pleasure having you on today and I look forward to speaking with you real soon.
Andrew Einhorn: Thanks, Brian. It was great being here. Appreciate the time.
Brian Thomas: Bye for now.
Andrew Einhorn Podcast Transcript. Listen to the audio on the guest’s podcast page.
