So welcome to the NFX podcast, and we'd like to welcome Alex Babin, the CEO of Hercules AI on today. Alex, thanks for joining us. Thank you for having me, James. It's a pleasure. Yeah. Well, you and I have known each other now for 7 years. We, we gave you a Pete seed check way back in 2017. You've been working on hercules since 2014, you're well ahead of the current boom in AI.
And so we thought it'd be great for our listeners to hear your thoughts about where I is today, and learn more about Hercules, AI, and why you're doing what you're doing. So, Alex, you just, quickly give us some background on you. I'm Alex Babin, CEO and co founder of, Hercules AI. I was born in Russia and moved to Bay Area about 12 years ago when my first, startup which was also in AI space, but, interactive video, object image object recognition.
My co founder, and I, we leave, in Campbell and San Jose, here in the area. My co founder, Givork, is also into AI, but he's been building before, joining 0. He's been, and that hercules, he's been building, skin cancer image detection algorithms. He probably saved more lives with his work than I ever met people in my life. And then we joined forces to build, Hercules AI to actually save time Pete people can spend it on the important stuff instead of, Pete tasks. Got it.
And how many, how many people are on the team right now? So right now, about ninety Pete, spread across multiple, countries, our headquarters here in Campbell, Silicon Valley. We have office in Canada, in New York, and, developing office in Armenia. Fantastic. So I guess, you know, you know, we met years years ago and and and it's recently you've changed the name of the company. Could you quickly tell us sort of the basics about what you do and then why you changed the name to Hercules.
Originally, a company called 0 systems. Which is a quite good name for technology company, and it was standing for 0 time wasted for Moge workers. But recently, not just our product offering expanded, but also the tasks, we build AI products for expanded dramatically. And some of those tasks are so hard to automate that at some point, we realized that the name doesn't reflect what we're doing. Herculean efforts, cleaning the audience Beller, and things like this, that's what AI is doing now.
And the reflection of the name change was basically because we're doing so many things that before we're considered impossible, now it has a broader meaning that Hercules can take care of things. Well, and and, you know, you you and I were actually on a whiteboard, I think, when the the name change came up, but, about a year and a half ago. Is that right? Yes. And it was your idea. You looked at our and you looked at our product and said, wait a second, guys.
You your platform called Hercules, Hercules AI, and you are not putting it in front of your customers, the the name, which is great name to describe what you're doing. Why you are hiding the great name behind behind the scenes, and we flipped it over. So now Hercules AI is both the name of the company and brand and the product itself. And I would tell you, James, it paid off nicely because people love it. Absolutely. Everyone loves Hercules. Everyone is neutral about 0. Because 0 is neutral.
Right? Well, that's great. I'm glad it's paying off. It, it, it, it makes life more fun to have a great name as well. So let's let's turn it back to 20 14. Well, before the latest AI moment that we're having now. So you're working on the idea of an AI model that could summarize long form texts into concise, easily consumable, like, snippets, right? Like saving people hours and boosting productivity. Where did you get that key insight? And you know, who are your first customers?
Yeah. So, actually, we were classic, approach of, the technology in the search of a problem because the problem itself was very challenging. And, again, it was pre LLM, pre transformer architecture, pre everything. It was when AI was still called ML, and NLP, not sexy, right?
Not like right now, but we encountered the problem when we realized that summer is a, like, the consumption of the text information takes the, like, consumes the most of the time from, the end users when they are processing information. And it doesn't matter what industry it is.
You might be a lawyer, you might be a financial analyst, you might be a reporter, it doesn't matter, consuming text, understanding it, summarizing it, and extracting valuable insight from it is essential part of every single workflow for knowledge workers. And we decided to attack this problem. And of course, it was pre LLM. We started building our RNN, models, but to make things even more challenging, we decided to build all the models working on edge.
Like it was not enough to build, complexity to build, a model that would work for general type of text. You were trying to build it on the edge, meaning, so that it would run on my smartphone Right. Without having to go back to a server for security reasons. Actually, not. Security came later. It was a byproduct. You know why we did it? We didn't have money to run it in the cloud. We were cheap. We said, wait a second.
Smartphones are powerful enough, to process those small models let's use the devices of our end users to actually do the processing. And then later it came back to us that wait a second, but it's also secure, and it allowed us to enter regulated industries where you cannot send the data outside of the security perimeter. And that was probably one of the most important things that happened accidentally.
I would love to claim it saying, well, we were smart enough to predict that in the future, you'll need to run LLMs on agent, on edge devices or inside security perimeter, but unfortunately, not. We were just trying to save money and build something to run on the client infrastructure of the hours. Got it. And so then you ended up with your first customers being in these regulated industries where security was really important.
And then you got the system to be more and more secure, more and more robust. Is that right? Correct. And once we made the first model work and it was still pre LLM, pre transformers, it was incredibly hard to do that. But we managed to do it and it worked, and we started looking for beachhead market because beachhead market in enterprise is critically important. I still do believe that it's impossible to build something generally available and applicable everywhere.
At least not at the very beginning. You have to focus, and you have to be finding this, beachhead market that you can tap into. For us, it became legal, large law firms. And that's where you and I met James. And I still remember pitching you, telling you, hey, we're building this automation, products for, legal vertical for large law firms. And you told me it was brave and stupid enough at the same time because, actually, it's one of the hardest market to Beller.
But at that moment, we already, got a couple of clients and they loved the product. And it still was incredibly hard to sell, but also had some benefits. And then later, of course, it expanded to much broader markets, but still inside regulated industries, insurance, financial services, and so on and so forth. Yeah. You know, you know, the, the legal tech 3, the legal customers have been really hard to sell into for the last 20, 25 years. They've been very slow to adopt new technologies.
That has changed in the last 3 or 4 years. And I think, you know, we've got other companies in in in the NFX Guild, you know, like Darrow and even up and and and whatnot that have been doing great with with legal tech. And, you've, you were a beneficiary of that, but you very quickly got out of just selling to legal folks, going to consulting companies and, and that kind of thing.
So, when, what would you tell yourself back in 2014 or 17 when you were cranking away building your own ML models and trying to get this sort of, basically, small version of James to work. What what would you tell yourself today? I would say one of the most important things that we realized, later than I wish we did is quality. Quality of enterprise products. It's not just related to legal.
Of course, legal, large law firms are very demanding, but every enterprise product, company is, is very demanding. And the approach move fast and break things doesn't work well. So we entered the market really fast with a half baked product and started getting clients. And I think we were absolutely lucky not to lose any of those clients because we substituted the, the lack of quality of the first product. It was like MVP, of course.
We substituted it with absolute obsession with a customer engagement quality. So, and that allowed us to not lose any clients, and we still have all the clients since 2017, 2018, we still have them with us, and now they're buying more products and they say that we are the best vendor they ever worked with. So I would say if I had to tell myself, back then what to do and what not to do, I would say slow down and spend more time on making the product better before trying to scale it.
Because we've been always under impression that Silicon Valley startups, no matter what you do, should scale incredibly fast. But that comes back and bite you in, in the back, later. And we only manage to survive that kind of a, tornado of global quality that was not still there by being obsessed with clients. And they gave us everything because of that.
Now when the products are on the next level, it's much easier to look back and say, Beller, we should have pay more attention to the engineering debt and not spend too much time building new feature and t features until we polish the existing ones. But, inside 2020, it's basically what I would say to myself. Got it. Got it. So let's talk about where you've evolved to today with Hercules AI.
So now you're a platform that allows, you know, the assembly of, what, lots of different applications in days or weeks, not months or years, and that lets you expand into more industries. Where are you today? And, who who are your customers and what problems are you solving for that? So let's start with what we actually do. So hercules' platform to assemble like an assembly line, right? In manufacturing, but to assemble virtual AI workers.
Let's unpack what virtual AI worker is, and then we'll we can, like, it will be obvious who we're selling it to and who is using it, who's benefiting it from it.
So virtually, a worker is basically a replica twin of a particular specialist might be a lawyer, but again, not doing the practice of law, doing a business of law, doing operations, everything that people don't like doing, or it might be financial analysts, analyzing prospectuses might be, operator doing private, private equity capital calls processing and so on and so forth.
So what we do believe in is that, humans doing hard jobs on minimal wage, while Roberts write poetry and paint pictures is not the future we all want. We want machines, AI, to take care of those mundane and Pete tasks. So people can actually focus on creative things. I know it goes a bit against the grain of January FYI as it is right now, which generates pictures and music and everything and it's fine.
For us, our mission is to build the AI powered workforce to actually enable people to spend more time on doing productive things. And it's been always from the very beginning of the company. It's always been the James. And if we look at the greatest companies, in recent years who emerged, on top of existing industries, for example, Uber, right? It's the largest taxi company in the world without having any cars on balance sheet.
We look at Airbnb, it's the biggest hotel chain in the world without having any property. We wanna Beller, we wanna build the largest workforce company in the world without having anyone on payroll. Because those are AI virtual workers, we give them to our clients to do the work that, typically people are doing. And as Sam Alpin was saying, there will be companies, bill into our companies that are run by a couple of people in army of AI Right robots. So that's exactly what we do.
But looking backwards, when we started, we we've been building on top of something that is very well hidden elephant in the room. There's a $1,000,000,000,000 problem on the market. I would call it invisible problem. Everyone knows about, but no one's talking about, talking about clunky old software. And organizations run on software. My BRP, my BCRM, might be anything. There are 1000 and south thousands of, pieces of software in each organization. These pieces are not connect connected.
People are jumping between them doing some manual work and so on so forth. So we build virtual workers to actually interconnect that software, but not just software, people as Beller, different departments. And that's exactly what virtual workers are. We started them building vertically 1 by 1, started in legal automating compliance, time capturing, quote to cash invoices, those things that actually consume a lot of time.
And then went into enterprise, broader price like insurance, financial services, and others. But we stayed inside the, boundaries of, regulated industries. Got it. And, aren't it feels like there's a ton of companies doing this now? How do you stand out? Of course. So there's so many companies that, want to make sure that people spend time on what's really important. They do it. They're all addressing it differently. Let's, imagine a spectrum, right, like the the line.
On one side, that would be RPA with a very simple processes to automate. Even if you add AI to it, it's still going to be AI powered, RPA. Robotic process automation. But on other side, there might be gigantic projects being taken care by, C3AI or Palantir, which are amazing companies, but their projects are gigantic. What we do we build, we've built Hercules as a platform as assembly line.
You can take components that already exist, and working and, and, test it with many customers and assemble the application, the virtual worker with our help, of course, assemble it pretty quickly and put it into works, to automate those very complex workflows. So we're in a basically sweet spot of intersection of very complex workflows automated by, AI virtual workers, which can be delivered really, really quickly. We're talking about Wix Not even months.
And so are you going into any particular types of companies, high security, mid security, large companies, medium sized companies, you going into companies with, like, a Snowflake implementation or Okta? Yeah. We focus on the large enterprises. Some of our customers are fortune 1000. Actually, most of them bigger enterprises. They all care about security. And also they are in, regulated industries. Again, financial services, insurance, and legal.
That means that we need to support all of the infrastructure they already have. And you can't build something like this to support, all of those, underlying infrastructure overnight. That what took us 7 years to build. Hercules has 200 components. For example, if a client using Snowflake and Okta we can support it. And the virtual, worker is gonna be delivered exactly, that that exactly feeds the enterprise, infrastructure. This is critically Morgan.
And another important thing for enterprises is cost of ownership. If they already have systems in place, They don't want to change those or rip and replace. It's important for the new technology, new products to work on top of what is exist already. And so what's the typical process? When you come into a company, you say, look, we're gonna give you additional workers we're gonna charge you a SaaS fee. How how do you charge folks? We charge $300,000 per AI worker per year.
It says, though unlike, many other AI companies, we are not hosting those, AI systems on our side. Remember, they are running on infrastructure of our clients. So our margins are incredibly high. If you compare with, other, companies where you need to spend enormous amount of money on GPUs and infrastructure. We don't have those costs. And also for clients, it gives predictability of how much it will cost. We don't have variable cost and clients have it inside the security perimeter.
So this is this is one of the most important things for clients right now because we've seen, clients prototyping themselves, using available tools on the Morgan, and they see that the cost running things, even on the sure it's astronomical. And sometimes they are saying, wait a second. I'll I'd better wait until it becomes cheaper, better and faster. Though it promises enormous benefits it's still far away from enterprise grade product. And I'll wait, and it's too expensive to run.
So I'll wait. But if they have an alternative, like, with us, they just can grab it, install it, and have it up and running in, in a matter of weeks. And then they know what exactly what their costs are gonna be. Exactly. It's very predictable. But there are there are other things that are critical Morgan. I've seen company startups that are jumping in and trying to build enterprise solution. Without thinking how much it will cost for a client.
So for example, we are using hybrid approach, GPU Pete. You can, virtual worker can use GPUs for specific tasks, but then CPUs kicks in and you can, you you can run smaller models on CPUs making it very, very inexpensive, basically using utilizing the existing infrastructure. So for clients, it's critical to know how much it will cost on at scale. Kinda and so you're using, one particular LLM Morgan you're using multiple? We have about 9 right now.
Some of them are smaller and those are hours and course, we use the best, the, the, the best in class like LAMA and Nystrel, because we we don't wanna train our own models. We fine tune them. We make them specialized. For example, we, just released the new model called Rosetta Stone. Why it's called Rosetta Stone, because it does something that no other model in the world is doing. It's structured data transformation. It's like ETL, but without rules.
It takes any structured data and turns it into another type of any structured data. It's just 7,000,000,000 parameters, but it beats GPT 4. By 30%. Got it. So this Rosetta Stone comes with your product when somebody becomes a client of Zeus. It's something you guys develop. Because you've developed your own ML models, which you're still using.
Yes. And now you're actually adding some of the other models and and Rosetta Stone is just another example of another one that you built internally or You're right. And there is also interesting thing, about quality of AI products, and it also, gets back to, what models are being used and not just what models. How they're being used. So we all know about hallucinations. Right?
It's, it's a pretty big problem, and more complex the workflow you want to apply LLM 2, more complex it is, more hallucinations you will have. And of course, there are techniques to battle it But actually, the best way we found is neurosymbolic AI. It's where LLM provides some output might be extraction, extract something from a document or, database or whatever. And then another LLM turns it into rules. But the rules are 100% accurate and then delivers those rules to another Beller.
And that's how you avoid error propagation when you build very complex systems. For example, Samawari, I, virtually a workers, they have 4 or 5 James working as Ensemble, one after another. And in order to remove this error propagation, you have to have neurosymbolic AI, and they had to build models that actually do that. It was pretty labor intensive, but now we can chain any models together and make sure that they are provide absolutely highest possible quality. Got it.
So this goes back to your discussion of how you've built out the infrastructure and the processes to really take care of the customers so that it's easy for them to implement without hallucinations. You you didn't actually come up with llama or with with, GPT 4, but you figured out how to structure them and implement them inside these large organizations. Yes. We believe James are they already commodity, and they're gonna be gonna be cheaper, faster, and better. And it's normal.
For example, right now, James, you don't care where electricity is coming, when you're making coffee in the morning in your coffee machine. You care about the quality of your coffee. And that's exactly the same gonna be happening with, LLMs.
They're gonna be So you're so you're predicting that the cost and the stickiness of a GPT 4 or 5 or 6 compared to an anthropic compared to a llama is gonna go to I would say it's gonna be so easy to switch between them if you have the infrastructure in place. So for example, I'll give you another example. Right now, let's say we deploy, we give our customer, AI application, a virtual worker that has, like, 5 models inside, and one of the models is Lambda 2.
Next year or this year, LAMA 3, being released, for example. How do you change? How choose swapped models? Without rebuilding the whole application from scratch. We learned that hard way 5 years ago before LLMs because we needed to update models. So we have a module, which could be a separate startup on its own, that does model swapping. And for example, reinforcement learning data stored in a separate database.
And when you replace the model, automatically, the reinforcement learning data being set up on top, and it the model Flint being fine tuned automatically without without us even touching it. So this is an example how infrastructure will actually drive, the way, models are gonna be used. So if it's so easy and painless to swap, LAMA 2 and put instead of an anthropic, Why not? It's about better, faster, and cheaper. Got it. Got it.
So all the people who invested in these big models are gonna lose all their money. I don't think so. Like, electricity generating companies are not they will turn into PG And E, right? And, as much as you don't like PG And E, but I think they're gonna be important part of the infrastructure underlying layer because you still need models. Right? But they're gonna be, they they gonna be come up. So you're charging 300,000 per virtual worker.
Can I have multiple virtual workers for that same 300, or do I have to pay 300 for every individual? One virtual worker can do the work of a thousands Pete, thousands of people. Right? So or, automate processes for thousands some, thousands of Pete. We separate those, virtual workers by type of tasks or specialization they focus on.
For example, you might have one virtual worker in, financial department who does quote to cash process, complete automation, but you can also have someone in legal department who is doing a payer Currier that does contract analysis and extraction from contracts and then passing that information to another virtual worker. That's the beauty. They actually talk to each other. We'll also build the system where virtual workers not isolated. They talk to each other passing information to each other.
But in this case, you'll have 2 workers and you'll pay 600 through how to teach. And we see clients who started with 1 and going basically, full scale with 3, 4, 5, and now putting much more, 1000000 of dollars in budgets. So for 300 k, I'm not buying a virtual worker. I'm buying a type of virtual Currier, and then we can let that scale up as much as I want. And then they can all inter intercommunicate. So so the within all these, so all this data is being passed between these models now.
I mean, data and security, and AI are such a hot topic. So why why is this integration of AI into enterprise business processes so scary for those enterprise? Number 1, because of the quality. Right? We all heard stories about, some engineer basically hacking, from hacking, chatbot for the, dealership company and buying, Chevrolet Tahoe for $1, right? So And things like this are inevitable when you have very generalized approach.
Basically, you give LLM to do whatever, whatever, you want it to do. But if you have a very specialized approach, the risk of having low quality is minimal. Or reduce to minimal. So one thing that enterprises all jump into, the use case is, of course, generated AI, let let's make it generate text. Let's do marketing emails, blogs, whatever. And that's great, but it's not, bedrock of those processes that those organizations organizations are running on.
So if you are fortune 1000 and you're using generative AI to, to generate marketing emails, well, it's still great improvements, but it's marginal. But if you are using generative AI solution to analyze if your financial institution, for example, analyze the fraudulent transactions and flag those and prevent those, this is where real ROI comes from. We've got data integrity issues. We've got compromised confidentiality.
You've you've got just a whole number of different types of security issues with implementing AI, and these organizations are just beginning to understand the implications of it. Is that right? Currier. And also startups who are trying to sell right now into enterprises, they will they still, will still have to have all those components to make sure it's, enterprise grade solution they're delivering.
Given that you've been doing this for 7 years, you've kind of worked through all of these these challenges. You've you've encountered them 4 or 5 years ago, found solutions for them. And now just part of the infrastructure that you're assigned to people as part of the platform. And we have to eat our own dog food before we build the platform. We have to build components and solutions ourselves to make sure that we deliver the highest possible quality.
And then we started turning those components into a platform. But we, like, looking backwards, if we started to build platform from scratch, we would not even know what to build, into it. What clients would need. We have to walk that that path to actually understand it and realize it. Right. It's interesting.
It's almost like, you started with a killer app, And then you built a platform on top of that, and that killer app was just basically reading long emails and summarizing them because that's what every knowledge worker does. And then and then you've expanded from there. Okay. And so 5 to 10 years from now, what's what's this whole AI market look like? And we've got Microsoft and Google, they wanna supply all this stuff to the enterprise. You're supplying stuff.
Okta is saying they're gonna apply, you know, add AI into their existing product. What what is how does the market evolve? I mean, all the listeners here are gonna wonder what your prognostication is. I'm a metaphor guy. I like metaphors because they explain things, visually. So I think of it as a as a planet. And the planet will have 2 poles. One Pete would be open source models, and another one would be closed source models.
Both will exist, and they will have the benefits, and, cons and pros but there will be continents of, I would say, two main pieces where AI first companies building something that Beller been done before. Because now they have ability to address use cases that without generative AI, were not possible to address. And that will be also AI enabled companies. Like traditional software, take Salesforce, they're investing a lot of into AI and doing it, for a long time.
Traditional businesses, traditional software that is AI enabled. And those 2 worlds would coexist. It doesn't matter that one will win over another. It's just different use cases they're gonna be addressing. And James as we know it, like, providers of James would be, again, commodity. So we will not even think what model we are using, inside the product. It the output, that we are using.
And right now, most of the money goes to, hardware providers because of requirements for GPUs and everything, but it it will also change. Models become Beller. They make, less they're less demanding for the, for GPUs. We use hybrid approach in our case, it's 10% of, usage of GPU and 90% of CPU and and the likes. So I would say everything is gonna be commoditized. But the platform platforms that will be able to deliver the ROI today end users or end customers gonna be actually winning.
Right now, it's opposite because it's very, I would say it's very immature market. Your prediction is that you guys are gonna be a platform, is there gonna be another platform, or will it just be Hercules? Of course, there's gonna be other platforms. It's obvious. It's a great, great way of, building something that never existed before. Because in this case, when you have a platform, your clients are telling you what they what they need. You don't have to go and kind of arrange them.
They will try find the use cases. They are coming. They're starting building things, and you just provide them with, ability to do that. That's where scalability comes from. So I would say it's absolutely obvious that a lot of companies will go that way, but what we Pete, even if they start now, it will take them time to build those components. Because it's so easy to prototype right now. You take a couple of models slap, blank chain on top, do rag, and it's still gonna be a prototype.
Clients don't want prototypes. I use the analogy all the time. We are at the dawn of aviation of AI. Right? It's a paper plane made out of dry, wood, dark, tape, and shit, and it's in the air. It's flying three hundred yards. Amazing achievement. We were walking on the ground. Now we're in the air. But enterprise clients need Boeing 777. With all the infrastructure and a cocktail served in the 1st class. That's what they need. And we are still, as an industry, pretty far away from it. Got it.
Got it. Got it. So what is going on, with, let's say, Google and, OpenAI in that battle? What they're they're what they're building now in terms of these James are gonna be subsumed to the platforms, but OpenAI has shown that they wanna build some sort of a marketplace, at least, that they're thinking about how they move up into that sort of platform area by letting other people build up locations. Is that something you're gonna see them doing, or it's just not in their DNA?
Well, they might, and it's pretty obvious idea, and it's great idea. They definitely will be going this direction, but I don't think they're gonna be going into hardcore enterprise because, again, that will require them building components, not just providing LLMs, or they will require developers to build those components. And it's still gonna be hard. I would bet on Microsoft because Microsoft has a lot of those components already built.
But regarding Google and OpenAI kind of a panel that we see right now. I think Google has a huge advantage. They have data. They have enormous amount of data, YouTube. They have Gmail. They have, Google Docs. The problem with Google, as I see it, it's not technology. They have smartest people in the world, and they have enormous amount of money. It's culture. Culture of innovation, and we know that culture eats strategy for breakfast, where Google is navy, Open AIs, like, they are pirates.
And I would bet always bet on pirates because they move faster. In this case, well, we'll see how it unfolds for Google, but so far it's so so interesting to see that battle unfolding from the sidelines and saying whatever who whoever whoever wins, we will be using whatever whatever is the best for our clients. 5 years from now, What are we gonna be looking back and and kicking ourselves from not realizing? That's a great question.
I would say we will probably be kicking ourselves for not realizing how, how still hard enterprise market is. And right now, we're kind of betting on, like, AI will change everything. The same way we were saying cloud will change everything. And the way before that, we were saying mobile will change everything. And internet before that. And it did, but we are underestimating how much time and efforts it will take.
So whenever we think like, okay, I have a great technology and a, in a product, and all those possibilities, I'll win the market in in 1 year. No way. It's still gonna be long and, and expensive. Okay. Well, Alex, this has been fantastic to talk with you. I, I think the fact that you've been at this now for for so long and, are so wise about how enterprise actually works and actually bringing AI to enterprise. You're like, to the leaders in in doing that.
It's just a real pleasure to hear your perspective on what's what's, been happening, what's gonna happen next. Thank you so much. Thank you, James. This was great pleasure.