Seeing the Market Differently: AI, Signal, and the Future of Investing with George Kailas, CEO of Prospero.ai

Welcome back to Financial Forward, where we go beyond the headlines to understand what's really shaping the future of consumer finance. This is season four, episode four. And today, we're stepping into the edge of what's next. Because while most of the market is reacting, some leaders are already seen around the corner.

ai is one of those companies. It's not just another fintech platform. It's a fundamentally different way of understanding markets, driven by data, signal, and precision. The kind of innovation that doesn't just participate in the market, it interprets it.

And leading that charge is George Kalis, CEO of Prospero.ai. George isn't building technology, he's building perspective. He sees patterns where others see noise, and he's helping investors and institutions rethink how decisions get made in an increasingly complex financial environment.

Today, we're going to unpack how Prospero works, what makes it different, and why the next generation of market intelligence may look nothing like the last. Let's get into it. Hey George, welcome. Thank you for coming in today.

Tell us a little bit about who you are and how you got to where you're at today. Thanks for having me. So uh I have a pretty long story in getting here, uh, but I will tell it as quick as I can. Uh I actually started investing when I was 13, uh, learned some important lessons during the tech bubble.

Uh, and you know, that made me want to start my education and investing pretty early. I worked for Bruce Greenwald at the Center for Value Investing at Columbia when I was 16, and my first value fund when I was 17. I actually taught myself accounting to get that job. Uh, and then I did some you know different jobs on the buy side.

I did a very interesting summer in 06 uh at Bear Stearns in Mortgage Finance. But I would say a lot of the lessons from my early career really led me to today, where, you know, even at uh small hedge funds, I kind of felt disadvantaged uh for a variety of reasons against the bigger guys. And, you know, a lot of you know what's in our platform today at Prospero came from, I think, learning, you know, lessons from you know what I might might have been missing, especially in the options markets.

Uh I got into AI pretty early, 15 years ago, started my first uh explicitly AI company. We're experimenting with some really cool technologies, evolving neural nets, where we had join IP with NYU, uh as well as uh some reinforcement learning before it was even called online learning. So we were really early to that stuff. Um but as I went through the AI as a service business, uh not only did I think we find some interesting ideas and how to use things like news and social media and other alternative data, uh I really uh you know kind of wanted to get back to my roots of you know building things that I think would level the playing field and help people more universally.

And that's really what we set out to do at Prospero. You know, we've we've been doing this for seven years. Uh the first three years, we were really just focused on kind of taking some of the advanced algorithms that we that we were dealing with with institutions and just simplifying them in a digestible way. And, you know, I was the first and really only user of Prospero for the first three years.

Um, and then we've as we've expanded, we've made it a little um, you know, a little easier to interpret. You know, I'm particularly proud of V4, which is the current version that people can use. Um, and then we also started writing a newsletter to prove that it worked because there were some step skeptics. Um, and that newsletter has done really well.

We've beaten the market by an average of 67% the last four years. We give those picks away for free. And so, yeah, we're uh we're an interesting platform where we have signals that could be more self-directed and we say kind of helps anyone improve as an investor or trader. Uh, and then we kind of teach how to read macro events.

There's always a big lesson in our newsletter every week, and then we teach people how to use the signals um to manage portfolios around those lessons. Absolutely, and I want to take you, I want to take you back. Um I want to take you back to the to the beginning. I want to take you back to the beginning or back a little while ago for a reason, and that is I want the I want the listeners to kind of hear how this came to be, which each listener will engage at a different point in time and or and or progression in your story.

Um and um, and it's important for us to hear each of those points. So I I really want to go back to the fact that you've been building predictive models since about 2011. And um what what what did you see in the markets that made you believe AI could actually do this better than traditional investing? I think there was a very important uh uh kind of like a seminal model that I built in my last full-time job.

I I built this model that actually could predict mortgages more accurately than than you know most people at the time. And I was using uh things like state laws uh to map them to really like foreclosure incentives, and that worked a lot better than what people were using at the time, um, with debt to income ratio was the main variable. Uh and and yeah, I think the move towards AI and wanting to have that technology was the understanding that like building those models were very time consuming for me and I could explore a lot of different data.

But you know, I I thought that the future would be, you know, me not having to read every state law myself to convert it to math. And and that technology had to be able to do that just as well or better than me uh at some point. So that was the the early bet that we made. Uh that that looks like a good bet.

So you've reported win rates around 58 to 61 percent compared to the S P 500. So, in plain terms, what are you doing differently that made your model an edge? This year's a little lower. Uh, we're at I think we're at 54.

This is a tough, tough year, uh, but still, you know, 54 is still real good. Um, so I think what we just do like really differently at our core, um, starts with uh I think what most people are trying to do, they're trying to trade and optimize in an automated way, um, or they're trying to predict price. And, you know, I had a lot of time learning that I kind of thought that was a fool's errand, especially if you're gonna say, hey, I'm gonna build a technology that can predict price better than a hedge fund.

Good luck. Uh so uh what we focused on instead was building variables that explained concepts. Like our most accurate, uh most used signal is called net option sentiment. And that, you know, very explicitly is not trying to predict anything, but does predict things often, as we've learned, gets out in front of news events all the time, whether it be at the stock or market level.

But all we were trying to do with that was explain behavior in the short-term options markets by doing things like isolating institutional behavior and clearing out retail behavior. Um, and then looking at, you know, we look at a lot of scaled variables, but we have linear formulas at the base, a lot of them, you know, that I worked on myself. Um, and so for something like net option sentiment, we use, you know, some uh a lot of you know price skew. So we're looking at like the differences between calls above where a stock is trading versus puts below.

And then we'll also look kind of similarly at open interest, but we'll look at like raw dollars behind something because we feel that that's important. But a really good way of how, you know, I mentioned I don't think it's smart to try to use AI to predict price or optimize a portfolio and give it give it that like whole claw. But the kinds of things we use AI for is originally volume wasn't in our net option sentiment. But we wanted to experiment with it more.

So we let our system say, you know, situationally, uh, if you can look at how you train and test and improve the model if you take a certain weight, then you could essentially earn more of that weight over time if you're correct about how you'll improve it, you being these scalers, these micro scalers. Um, and so we've seen on options expiration dates um as much as like 10 to 15 percent now. And we didn't teach it to do that, but that's very intuitive that volume would be more important on options expiration dates.

And I say all that because now I'm gonna circle back to your question, which is what we do is not necessarily about any of the single signals alone. We built these 10 signals to be a language that makes it easier to interpret the markets. So that's what I think when I make the portfolio, and now I'm training some other people to be the portfolio managers because my time is getting a little more limited. Um, but it's about the heavy lifting that the signals do, and I think the asymmetric information that's being presented because of the different way we approach the problem that then enables you know people to be smarter.

Like one of my favorite stories with Prospero, a former reader came. Um he's a he's he used to be a pastor, pastor of a mega church and he's a PhD in theology, um, not the the profile that you would expect to do well. He now is, you know, the main writer of the weekend letter. And in the last two years, his portfolio is up 400%.

He's doing better than me. Um, and that kind of speaks to what I'm saying, where you know, there's signals, there's a process that we teach. Um, and I think that's the right way to look at this because, you know, one of the things that I always tell people, and I think it's always a fair question in a similar way, that you know, you asked it when people asked me to evaluate um, you know, what this XYZ wealth management tool does. I'm always just like, well, it's hard to say.

They're not being very descriptive. I'd have to like know exactly what was going on in their code. But beyond that, I'd ask you this this company hasn't even raised a hundred million dollars. You know, Bridgewater could lose that in a day.

Uh, what exactly do you think that they have that can compete head to head in that? So that's what we really do differently. We're not trying to compete head to head with algorithms trading against each other. We're looking a lot of at teaching people how to use these signals to read the directionality of things like what algorithms are trading in in the market and getting better at interpreting that themselves.

And that's a huge difference. Absolutely. And and it's it's paying off, obviously, right? So I I I I we specialize in an AI tool that does something similar for a different reason.

It's not focused in the investment industry, but the synthesis, the synthesizing of the right data in order to understand the bigger picture is so important. But it's who's doing that synthesizing and what data they're taking in order to determine what that outcome is, right? So anyone can can build a model, but it's the brains behind who built it that is actually defining the outcome of it. And clearly you have a lot of fingerprints around the backside of this model and how it was put together and how effective it works.

So hats off to you. And that's that's just terrific. But so talk to me about these you know, these AI-powered tools are are great for institutions, they're great for big buyers. But how do we get this down to the everyday investor?

And how can they access something either this or something just as effective as this? Yeah, I mean, I I'm honestly really scared of the reckoning that's coming when there's a bear market because I think a lot of people are using these LLM tools, and I like everything that I've seen in terms of how they recommend things, it does not seem to it does not seem to be that they understand um risk very well. Um and you know, we are even experimenting with our alerts product um that is a combination of our signals and um and LLMs.

And and honestly, we're reevaluating, especially in the last few weeks, how much we even want to lean on the LLM part of it. Um, because they've just been giving really like during these last couple months, they've just been giving really, really bad um results. And they don't, you know, it's because I don't think they can read the market that well. So I mean, I think, you know, I feel very strongly that um that the more automation involved, the less valuable the tool actually is gonna be to someone, right?

Because like, let's talk about all of these agentic systems. It's the same type of thing. Like, even if all things are equal and your agent is just as good as Citadel's agent, which I really don't believe on its face, like they're still getting better execution than you. So I think the more you're just like handing something off to an AI, the more you're really just, you know, handing yourself um a kind of ingrained disadvantage that can't be changed.

Um, but you know, one of the things that I like, you know, I really think that's that's huge in the market that's changed is like you look at what happened with like Palantir and Tesla recently. Um that's really uh retail investors showing the power that they have. Um, and you know, there was a lot of negative articles about the way that they were valued, but at the end of the day, enough people continued to see value in those companies and they kept on going up, and the market's about agreed upon value, no matter how much they might be overvalued by traditional standards.

Um, and I think that's a really good example about how uh, you know, it's a it's a lot of what the hedgefront community does, like good research, accreed agreed upon value, a community built around those things. Like that's where I think the real advantages are. And also using, you know, good AI tools to support that research, I think is a great way. And there's a lot of good tools out there.

But I think the more anybody finds themselves saying, like, oh, I'm gonna open up OpenClaw on a perplexity computer and make my millions, um I I think that there's gonna be a lot of downside in that. All right, financial forward listeners, let me tell you about someone who's redefining what it means to make an impact in financial PR, Angela Nibs and her powerhouse team at Maven Communications. I've worked with Angela personally, and I can tell you this isn't just another PR agency.

Maven is about people, their experience, their creativity, and their relentless drive to deliver results. Whether it's launching products, amplifying your brand's voice, or establishing thought leadership in the financial world. Angela and her team at Maven don't just play the game, they set the pace. What sets Maven apart is their understanding of the industry's complexities and their ability to turn those into opportunities.

They don't just connect you to the right people, they elevate your voice so you lead the conversation. This isn't cookie cutter PR. It's strategy with heart, innovation, and grip. If you are ready to level up your visibility, strengthen your reputation, or lead your industry into 2025, you need to connect with Angela and the team at Maven.

Visit Maven PR.com and schedule a free consultation. Trust me, it'll be the smartest move you make this year. You were talking about um, you know, AI is the great equalizer, right?

And and I I get and believe that if if it's used properly, like what you're doing, and and um um um using a holistic approach to using AI in the right place at the right time for the right reason, right? And you also describe something where you're worried about the next bear market and some of the tools that are being used today, and how that's going to amplify or compound a bear market and and its velocity and and maybe even its its the length of time. Um what is um what needs to be in place to make sure it doesn't actually make markets move more volatile or unfairly?

Like what are there controls that need to be put in place? What what are your thoughts about that? I mean, I'll even go to the the current market if something comes out uh in the news about the the um the oil industry in a negative way, maybe in the Middle East, um it affects the markets immediately and it almost makes it look in the volumes the market moves, it almost makes it look like they're all attached to some type of intelligence that is absorbing the news instantaneously.

So speak to that and then tell me what kind of controls you think need to be in place or or do they. So I mean, I think there's a lot of like a lot of controls that that need to be in place with AI that just aren't gonna aren't gonna happen. Uh and and I think there's a realistic reason for it where it's just like there is kind of an AI arms race, and and I think there's a real hesitation, and and there's reason for it, but I think there could be good regulations. But I actually think um in terms of like institutions using it, trading, integrating information, really, I just think it's a long shot.

I think it's a long shot to get anything and a really, really challenging thing um to regulate as well. And I think that's that's part of it. But I think, you know, I think I could think of a really good example to uh to kind of represent, I think the choice that people have. And I'll just talk about Claude, right?

Because it's a choice of how far behind you might be on those things. And it's just like if you're saying using Claude to say, like, you know, what's the best stock to buy right now? Or or you know, trade oil futures for me, right? I think that kind of thing, you're gonna be behind all the robots that have more money, have a bigger connection speed, have better educ execution, right?

I think that's a given that people have to accept. Like, but there's another side where it's just like Claude is uh, I teach a lot of people how to uh invest or trade one-on-one and I recommend things for them. Um, sometimes it's not using Prospero. Sometimes I'm just like, for the amount of time you have, I think ETFs are for you.

Um, but that has completely changed in terms of how I do it recently. Like what I used to do is I'd say, like, here, you know, you can I think start with a watch list, you can start with Prosperos, you can start with just like a list of stocks that you're interested in. And then I would read, you know, I'd recommend Barons as one of the sources, and I'd say, you know, look at these numbers. And I would give them basically a list of things to read and look at.

Um, that's how I used to teach. Um, as of six months ago, um, maybe, maybe a little later, but roughly, I would say, okay, now just take a list of stocks, put it in. Here's a prompt or two to run for Claude to compare the stocks. And then I would say, whatever comes up, you know, if you don't know what a what a PE ratio is, ask what a PE ratio is and and get that answer, right?

If you don't know how to situate, you know, the PE ratios amongst each other, you know, ask about that. And then ask about, you know, maybe how how projections might look, maybe some of the assumptions, like, you know, the kind of um, you know, margins or expenses that would go into that. And that's completely different. People are learning so much faster because, like, not only is it more structured, it's a companion that I think is really helpful to people.

And I think that that is the real bifurcation right there. If you are trying to get AI to do everything for you, then AI has the same problems as most financial products that I've ever seen. The ones that say all you have to use is this one number and we'll make money for you. That problem has existed forever.

I see you smiling. I know you see that too. But like what is easier now with AI is you can actually learn faster. And the best way to make money was always learning.

You could just do it faster now. Right. And if you look at how we acquired that learn faster in the past, you look at the companies that learn the fastest. Let's look at KPMG, let's look at Booz Allen, let's look at uh PwC.

Let's look look at those top companies that can move the quickest and are the most agile. And they've been that way historically. Because they've been able to acquire the top talent through the university system in the in the nation, ensuring themselves that knowledge gap or filling that knowledge gap that you're referring to and having access to the quickest information, being able to process it as quick as possible with the best resources you could. Now you replace that today, but you're not replacing the whole funnel from end to end, you're replacing or augmenting how that information is being processed.

So I I agree with you 100%. But so let's let's say looking ahead five years, uh where where do you see specifically your industry? Where do you see this going? And will AI just help people make better decisions or will it start making decisions on its own?

So I mean, I I think the direction is going to be more and more automated, um, for sure, because that's what people want. And I think, but you know, I think the other side to that is um, you know, whether it be hedge funds that are doing that or people that are using those services, I think there's gonna be, I actually think there's gonna be a convergence and an evening out of those returns, right? Where they're all kind of like finding the same alpha. And I actually think it'll converge towards the market even more.

Um, but I think where, like whether it's on the institutional or retail side, and this is like, you know, kind of why I continue to advocate for it. Um, I think if you're looking on like the institutional side, I think people that are going to be able to um look at more like, you know, kind of like what we're doing, except we call them signals, you know, more factor-based models. They look at like reducing the data set to be able to find more connections like that. Um, I think that kind of thing, any kind of knowledge engineering, I think there's still like a lot of advantages in that.

Um, but there's no question on the retail side that you know these automation products, I mean, I think there's probably gonna be kits where you can order like a pre-programmed open claw computer that you just plug in and it's gonna like start making you money like trading, you know. Maybe like I think we're already seeing that in the prediction markets. I mean, I think that's gonna expand. Um, in addition to in addition to these, I mean, we're moving towards 24 hour trading.

So I think a lot of people are gonna be sold on this idea of like, hey, I'm just gonna plug this in um and I'm gonna make money. And I do actually think um that is an upgrade for a lot of like what is it, like 95 to 90, nine 95 to 90 percent of retail investors don't beat the market. Um, I think that'll probably be similar with these automated systems, but like we're talking about, I do think the heavy losses will condense. Like I think you'll have more reliable gains from these automated systems because I think they'll be good enough to you know get closer to that market return.

Um, because if you look at like, you know, what makes retail investors unsuccessful, um it's the emotion. And then it's also just like, you know, the blackjack problem that I always like to tell people, which is that like people make the retail investors make the same uh mistake as blackjack players, which is that when they win, they'll bet more. And you know, when they lose, they'll pull back. When it's like the game theory is the reverse of that.

Um, so I think a lot of that will be taken away by the by the automations, but like I said, I think there's that much more opportunity in that situation for you know communities that do good research that buy stocks together. Because in those situations, and you know, this is something that Prospero is betting on a lot, like in those situations, you know, where it's a bunch of robots, right? Who who leads the rope, the momentum robots, right? If retail moves in size and all these bots are looking for momentum, you'll actually have retail at the base of a good amount of these trades if they can form these communities.

Right. And so talk to me about talk to me about who your customer is, your optimal customer, and and tell me what you would want to say to them. Um, so we have we have two primary personas. Um the first is, you know, kind of I would say like the the reformed meme stock trader, and they're like, you know, late 20s, early 30s.

Um, and you know, a lot of them have just gotten burned on, you know, those kinds of just pure social plays, and they're looking for more um data. And you know, to those people, I always say, like, come look at, you know, start to look at your investments before you make a decision, check out the Prospero signals, especially in that option sentiment. Check out the market level ones, SPY, QQ, Q, net option sentiment. See if it makes you better, see if you improve.

Um, I've never had anyone that I said that to come back and say that they didn't um improve. And then I would say, um, we our other main persona is like kind of like the um, I would say long-term investor that's late 30s, early 40s, heard about AI, wants to improve their process. You know, we see them doing some kind of like what we would say enhanced dollar cost averaging, where they like look at our signals um for what what their you know big names in their portfolio and like kind of allocate accordingly.

And for you know, that it's even easier to measure. I would say, you know, take a take a sample, take, you know, split your portfolio in in you know, 50-50, do you know some of the enhanced dollar cost averaging, doing some of the traditional way, see what does better. And where do they get you? Um, prospero, uh, just go to prospero.

ai. Um, or if you're curious about our products, uh, you can email me at george at prospero.ai, and I always love to give out you know free trials, help people with their learning curve. Thank you very much, George, for being here.

I appreciate you. Thank you for having me. That was George Kalas, CEO of Prospero.ai, and a conversation that really underscores where this market is heading.

If there's one takeaway, it's this the advantage is shifting. It's no longer just about access to information, it's about the ability to interpret it, synthesize it, and act on it with precision. Prospero is pushing that boundary. And George is one of those leaders who isn't waiting for the future to arrive.

He's actively building it. Hey, if you found this conversation valuable, share it. Send it to someone who's trying to understand where markets, data, and decision making are going next. And as always, thank you for being part of Financial Forward, where we focus on what matters, what's changing, and what's coming next.