Key takeaways
- Radu expects the AI hype to cool soon, but for the gold diggers, not the tool makers: companies built on an ERP in two nights or on AI employees will fade, while the tooling enterprises need is the real opportunity.
- Inside most enterprises people already use AI without the organization knowing, and IT has no tools to audit, procure or control its access and cost. An admin account in a person’s hands can be audited; in AI’s hands, it reads everything.
- Data is what sets a bank or insurer apart, since processes can be copied and people hired away. Share that know-how with a frontier model and everyone learns your “black shoes” pattern, so it has to stay yours, in a model focused on your vertical, wherever it runs.
- What an enterprise can deploy now is a connecting layer: one entry, exit and audit point between today’s models and legacy systems it will never take out, with access limited in time and scope. That is where Radu sees FlowX.AI.
From an AutoCAD library to Microsoft
Mike
Welcome. I’m your host Mike Parsons and we are here once again in the FlowX.AI headquarters. We’re having another conversation where we’re going to get into mission critical enterprise technology, but I want to mix it up a little bit this time. We’re going to expand our horizons. We’re going to get the brain cells working. Because today we’re going to get into what the world of technology looks like. We’re going to look a little bit big picture here. We’re going to look at the patterns and trends over the last few decades. And what we’re really on a mission to do is to have a conversation, to go deep with an expert, such that we can connect the dots and figure out what’s happening now and what’s happening. Next. And to do that, I am so happy to welcome Radu Georgescu, an advisor to FlowX.AI, someone that’s going to help us accelerate mission-critical technology in the enterprise. Radu, welcome.
Radu
Hey, thank you for having me. It’s amazing to be here.
Mike
It’s great to have a fellow global citizen like myself. You have done so much. I feel like you and I could talk about the good old days and which bubble, and which cycle, and where were you when that happened? You’ve been in the venturing technology space for more than thirty years.
Radu
I’m really old, man.
Mike
And I’m feeling like I’m right there with you. I would love you to give us a sense of some of the breadth and the spread of the things you’ve done. From hardware to software, from Europe to North America, from starting your own companies, investing in other companies, advising companies. Give us a little bit of a snapshot of the spectrum for what I do.
Radu
How much time do we have?
Mike
Plenty.
Radu
So I’m a mechanical engineer by training, by education. And immediately when I finished university, graduated university, that was back in 92 here in Romania, I sold my graduation project, which was an Autodesk, AutoCAD library, and I sold it to the Autodesk distributor in Romania. And that was amazing. I was, what, 21, 22 years old. Learned a lot, started. The company doing more things, and one of them was an antivirus.
Mike
So hang on, you were doing antivirus in like 94, right?
Radu
There were like four antiviruses in the world, and an antivirus conference at that time was with Eugene Kaspersky, John McAfee, Bogdan Dimitrov, and myself. And that was kind of an antivirus conference of the world.
Mike
What was it like doing antivirus in 94? I’m curious, like, what were the viruses that you had to defend against then?
Radu
Maybe 50.
Mike
Is that all?
Radu
And were on discs.
Mike
Yeah.
Radu
And so amazing, amazing days. Absolutely amazing days. Selling them online and th then I got into payments and that’s a different time of my life. But selling an antivirus online, selling a software online, no people were pushing pay and then an email would come to my email and with clear letters, was the name and the credit card number and everything. I was copying this with a pen and on paper, go to the bank, and please credit this credit card. So that was payment security online at that time.
Mike
I think that’s where you got the idea for your second venture in payment security.
Radu
So then the antivirus company got acquired by Microsoft in 2000. One, I guess. And that was the core of their security business going forward.
Mike
So this is amazing. So you just finished university with your AutoCAD template exit. And then the very next thing you do, you ended up selling it to Microsoft.
Radu
The antivirus.
Mike
Yeah. What was it? That must have been like a fairy tale.
Radu
Absolutely. It feels like, even now, looking backwards, it does feel like a fairy tale. Yeah. Amazing team, amazing. May I call them kids? I was a kid myself, but they were younger than me, creating the technology and everything, selling back in 95, 94, 95, 96 online worldwide, antivirus technology, and then Microsoft got interested and acquired us. And it was very early days. Microsoft was not acquisitive at that time. I think we were the first acquisition outside US and the fifth-ish acquisition ever in Microsoft. So very amazingly interesting, very interesting life.
Mike
So you would have been pretty young when that happened. Was that 32-ish? Yeah, was it stressful, complex, hard? What was it actually like, the transaction?
Radu
Exciting.
Mike
Really?
Radu
Absolutely. Absolutely exciting. Microsoft was like an amazing company at that time. Not that now it’s not, but I mean, it is. But very a lot of things I’ve learned. Most important for me, I’ve learned to be, as you very well said, you know, a global citizen and understand what the global things mean. I learned as well a very important thing that helps me even today: that there is what happens in Silicon valley today is going to happen in Romania one year down the line. So going looking at both worlds, you kind of see the future and you can arbitrage this future and create value. So that’s a huge opportunity.
Mike
Yes. So by your global point of view, you get this unfair advantage. You’ve seen the future.
Radu
Interesting, huh?
Mike
Maybe we should keep it a secret.
Radu
Yeah, yeah, yeah.
Mike
Don’t tell people. We won’t record this video.
Ahead of the wave, and knowing when to give up
Radu
And then I got into payments. Yes. So created two payments companies, ePayment. Very local here for all products, actually started the the payments industry in Romania, acquired by Naspers. It became PayU. And then payments company that was global but only for software by the name of Avangate moved it to San Francisco, acquired by Francisco Partners eventually. That was back in twenty eleven, twenty twelve ish. That was a very interesting project to actually relocate a full a complete company into the Bay area and sell it to a proper [unclear] in San Francisco.
Mike
Was that also still in your excitement category? Or was that a little bit more complex?
Radu
Oh, no, no, absolutely excitement. Look, I love being one step ahead the wave. So security before it was interesting, payments before they were interesting. So started companies before the wave, some of them successful, some of them actually not. So I got into blockchain back in 2011. That was way too early. And I didn’t want to get into Bitcoin trading and so on, but in technology. So our company was supposed to be what Ethereum became, but it was way too early. And that was a complete learning experience, so to say.
Mike
What’s the one thing, if I was to ask you now, that was unique in that learning, when you talk about it being a learning experience, what’s the one big thing?
Radu
Being a what experience?
Mike
Being a learning experience.
Radu
A learning experience. Yeah, yeah, yeah. Yeah, so it’s a very nice way to call a failure. So no, no, you learn how to fail, you learn how to fail graciously, you learn how to not fail twice in the same way, doing the same mistakes, so to be innovative in the mistakes you are doing. And that’s exciting as well. The most exciting thing for me is being part of a growing nascent industry. So to be able to not perfectly execute something that somebody before actually wrote the book about, but to actually write the book about what’s people are going to execute in the future.
Mike
And as you’re speaking, I get this sense that you have this very strong archetype of being a pioneer. You like to be trailblazing, but also great optimism that goes along with that.
Radu
I have to be. I think it comes with this because you. Cannot when you start things, you are getting a lot of problems. I mean, you need to solve problems, you need to invent new ways of doing things to figure out what the customer wants, what the customer needs. Even when the customer knows what they need or right. And it can be painful, it can be hard, but it’s as well so exciting. And that that’s the amazing part.
Mike
So I think if you know you’re fighting for something that really fuels you, compels you, it’s your purpose, then when you’re a bit stuck in the mud, that’s okay. You’ll keep going, you won’t give up.
Radu
Absolutely. And that’s the interesting part. People always say, you know, don’t give up, don’t give up. I think one of the important learnings that I’ve taken out of this journey is you have to know when to give up.
Mike
Yes.
Radu
Right? Because being persistent and not giving up can actually dig you in a hole that’s bigger than what you want to be.
Mike
And may I suggest that it’s when you stop learning that’s the signal?
Radu
Absolutely. I absolutely 100%.
Mike
Yeah.
Radu
100%.
Mike
Yeah. Well, what a great context for us to dig into some hot topics.
Radu
There are some more.
Gold diggers and tool makers
Mike
So let’s see if we can sprinkle some more in these hot topics. The first thing I’d love to discuss with you and get your thoughts on is, you know, I’ve kind of set it up earlier. Like, I want to talk about where we are now and what’s next. And I have in my mind, you know, the Gartner hype cycle where it’s like it builds up and there’s this peak of inflated expectation, and there’s a valley of darkness, and then we sort of come out.
Radu
Maturity, yeah.
Mike
It always reminds me this moment that we’re in of I was enthusiastically moving to Amsterdam in April of 2000 to do a startup and we had raised some money at the back end of 99 and like I’d fully relocated to Amsterdam in April. The only catch, Radu, is that in March of 2000, the NASDAQ had lost 15% in one day. And so, anyway, that quickly showed me that the peak became the valley of darkness. And I’m always reminded of that. So, it brings me to this cycle now. And I would love to get your thoughts on. What cycle are we in from your perspective? Where are we right now? And what do you see playing out of the future? What are your thoughts?
Radu
I think the obvious topic here is AI, what we’re speaking about. And looking at the cycle of the AI, I think we’re still growing. We’re still growing to the top and actually pretty close to going down. AI was nothing for the regular people until two, three, four years ago. Like really nothing. It was only in the mind of a few. And then it was so exciting to be able to speak with somebody that is not a person and get words, not information, right? And to get a partner of discussion that’s artificial. And so it’s so exciting. So the growth was so abrupt and so aggressive and so big. I do believe it’s going to blow rather soon. But it’s still growing. It’s still there. I think even and there are. I I would compare it with with the everyday’s metaphor of the gold diggers versus the tool creators, right? And the gold diggers are the AI people. So I believe that that that’s gonna kind of go down rather soon. But the tool creators are there to stay because AI is there to stay, right?
Mike
So give me an example of who’s at risk and who’s the winner here, to use your metaphor, the diggers and the tools.
Radu
Yeah. So I think that the diggers are companies that are trying to use AI to create an ERP, right? I mean, come on. I’m excited to be able to create an ERP, you know, the two of us together in two nights.
Mike
Yeah. The vibe coders.
Radu
Yeah, exactly. Co compared to 20 years ago, it was like 100 people 10 years, right? That’s amazing, but that’s not going to last in Norway. I think that AI employees is not going to last like an employee that’s going to be AI. I think that. AI companies like a founder and then a lot of AI for marketing for this, for that, for this, for that. I’ve seen about 50 companies like this, none of them being successful, to be honest. Yeah, I don’t think it’s going something. So, all these things together, I think, is going to bring a lot of bad vibe that’s going to go kind of down. But again, I haven’t seen the future here, so I what’s so so?
Mike
Let’s just draw an interesting comparison because I remember the promises of 99 and 1998, and we were talking about this internet revolution. This is even pre-cloud. Yeah, and what I thought was interesting, and I remember two thousand through two thousand three were dark times for anyone in the internet industry. I remember seeing commentators questioning, well, maybe this internet is just a fad and it goes away completely. Now you just think about today the ubiquity of the internet. And it makes me reflect. I remember in the late two thousands thinking to myself, all the promises that were being made in like ninety nine, two thousand, they were all promised that we’re going to have all this super high fast internet by 2004 or five. It all happened. It just happened in 2008 and 2010. It was a bit slower. And this is this natural part of the hype. So I’m interested, the picture you just painted with this one-man company with 10 agents fulfilling the other roles. It does sound a stretch too far, too early now. But could you see a world in the midterm where that actually does start to happen?
Radu
We have to define what midterm is. Midterm, you know, 20 years ago was 10 years, midterm today is like six months. So that’s one question. But before this, I think. For this to happen, there’s a lot of tooling that needs to be done. And I think this tooling is what’s the actual opportunity for success today.
Mike
Okay, so this brings us back to where we were before.
Radu
Yeah.
The tools enterprises don’t have yet
Mike
The tooling. Give me an example of a tool, a couple of tools you look at and think they’re winning.
Radu
Look, look, m my obvious investment case today, if I were to invest, is tool for AI infrastructure for enterprises, right? Wh we we’re in a point where enterprises, and we can dig into this whenever, but enterprises, you know, banks and like the big guys, right? You speak to their CISO or VPIT or whatever they want to call it. And they’re so happy. We have no AI problem. We are the perfect bank. We are the perfect this. And then you go in the evening for drinks. They start literally crying for where they are. Everybody in the enterprise is using AI, mostly without the knowledge. Of the corporate, of the corporation, right? It’s exactly where the internet used to be, where the credit cards used to be, right? What the hell is happening? And they don’t know. And AI is something that, so you open a username and a password, even an admin username and a password, to a person, that person you can audit that account. You can whatever they can do with that is pretty limited in scope and time, right? You open an admin account to AI, you are done.
Mike
Yes.
Radu
I mean, you are done. Yeah. Both for scope, for time. For imagination, for understanding and everything. A person can read whatever they can. AI reads everything, right? So that’s amazing. And as an IT department or whatever it is in a corporation, you have right now no tools to audit. To understand, to procure, to do anything about the access and the cost and everything about AI.
Mike
What’s really nice about the picture you’re painting there is when you, as soon as you go to enterprise, you get the scale, the volume, and complexity, which I think. AI and agentic and agent-based solutions love those situations. So we’ll get to there in a second. I want to press you though. I want to find out from you who do you think is where do you look at people like Palantir who are obviously working at that sort of level? Would you benchmark them as some of the toolmakers that are winning?
Radu
They should, yeah. Absolutely.
Mike
Who else comes to mind? Like, who do you benchmark? I’m really trying to understand the characteristics.
Radu
So, I don’t think it’s clear how this problem is going to get solved. So, Palantir is trying to solve in one way. Databricks is trying to solve from a completely different direction. And Databricks is. A potential winner, right? Then you know the Oktas and the companies that know your name and the access management. The identity and access management companies have an angle as well. Very interestingly. Payments companies have an angle for this.
Mike
Well, we’re seeing Stripe making some acquisitions.
Radu
Absolutely. And some declarations as well saying that we believe that payment companies are best positioned to actually create the infrastructure for AI for enterprise. Now let’s who’s going to be winning? You know, that’s a very interesting series to get popcorn and look at.
Mike
Absolutely. And I just want to go back to why payment companies are quite interesting. Obviously, you’ve had some success in this area. But what I see is high volume of very similar patterns, transactions. This is something that AI can work with very nicely.
Radu
Absolutely. Look, at the end of the day, AI is token-based. Token is some type of a money type of a thing, right? So payments manage money. They can easily manage tokens and understand where tokens are going and then put a layer of real infrastructure on top of it. And here they are. They know what they’re doing. So having said this, the other angle is to get something AI, like, so payments are payments, you know, access management, Palm access management, it’s one thing, and then Databricks is the data, and then it’s another thing. Palantir, it’s another thing. So I guess the other one is actually doing something new, right? Because you get payments and stretch them into becoming AI infrastructure. You get data and stretch it to becoming AI infrastructure. You get people on Palantir and stretch them to become AI infrastructure. Why not build why not take infrastructure for God’s sake and make it AI infrastructure? Right? So there there are so many opport opportunities. My bet is and that’s why I’m very conscious with my time and choosing to help FlowX.AI it’s something that is important for me. My bet for this is that infrastructure actually has to become AI infrastructure, right? So I think that the way that comes into play has a very solid choice have a chance very sorry chance to become the right infrastructure for enterprises for AI.
After the peak: layers and verticals
Mike
Well, this brings us to our next topic, which is you’ve painted this picture of where we are. We know we’re on the sort of sharp end of the peak of inflation. We know that in the future, there will be some rebalancing in the market. There’s too much hype and certainly there’s been a lot of investment.
Radu
And sorry, going off the hype hill doesn’t mean that the technology is not going to get used anymore or anything like this. It’s just that the companies are not going to be priced as much. And some people are going to lose money. Some investors are going to lose money.
Mike
Yes, I firmly believe, though, that we could see a very real situation that a lot of the value that is high at the moment in the market is expecting a very large total addressable market to be realized in a very short amount of time. I think it’s that short amount of time thing could move. I think what we’re seeing from this whole series is there’s real difference being made in the enterprise with AI already. So there’s a lot of evidence of its potential and the use cases. Absolutely. But the thing is, we’re interfacing with mission critical, human in the loop. There’s a lot of variables here that mean you can’t get on the fast track and steamroll your way through this. It takes time. I mean, I think about other cycles we’ve seen that have taken time, and they invariably, with exponential technology like AI, it happens, but it often happens in a slightly different way, shape, manner, and form than what we might.
Radu
Marginally subject to it.
Mike
Yeah, yeah. And I think that’s what you’ve pointed out. We’re figuring out exactly what that looks like. So let’s jump into the component parts of AI for a moment. I want you to imagine people often call this the full AI stack. So you’ve got like the infrastructure here. Then you’ve got the data. You’ve got the model itself. Then the workflow and orchestration. What you were starting to map out for us is potential places to win. So I want to kind of zoom in on this picture. How do you see the AI stack? Particularly when we look at the enterprise and where are all the opportunities? Do you see opportunities in everywhere from the orchestration right through down to infrastructure? Or is it weighted to particular parts of this stack?
Radu
I think that to start with, the layers are as you explain them, but then there are verticals.
Mike
Oh, okay, yeah.
Radu
And I believe that AI that is good at everything, like Claude or ChatGPT or whatever, are not necessarily here to stay. Or, if they were to stay, they need to become very focused on an industry. I think they will not afford, cost-wise, to know everything. So, to be able to answer your question: what shoes do I need today? They need to have a model for shoes so they can answer much cheaper and everything to you. So, I think there’s going to be a I don’t know what the verticals might be, but there’s going to be a banking vertical in front of whatever, right? And I think that’s an important part of it. The second part of it is that this know-how, the vertical needs to be able to become and it’s wrongly said, but let’s call it on premise. It’s not going to be on premise per se. But it has to become property of bank X and Y.
Why sovereign: the black shoes
Mike
Okay, this now you’re touching on a really hot spot. Sovereign open local models and AI. This is definitely huge. What’s you have you obviously have a high conviction that it’s going that direction. What’s driving the move towards that? Why are the frontier models a little bit at risk here? And why is sovereign local AI so hot right now? What is it in the enterprise that makes that a good fit?
Radu
At the end of the day, if I’m a company, I’m a bank or insurance company or whatever, like really whatever, what makes me different than you, the other company, competition across the street, right? At the end of the day, it’s basically data that I have. That’s the biggest clear differentiator. Processes can be copied because you can see them. People can be stolen.
Mike
Okay.
Radu
And so on. But data is actually what I for example.
Mike
That customer and our customer relationship, their details, their history, their context.
Radu
Absolutely. So I can go to that customer and sell him black shoes, and I know he’s going to buy it.
Mike
He just bought 10 black shoes.
Radu
You are going to sell him red shoes. He’s not going to buy it.
Mike
So he’s my control.
Radu
If I put my know-how about the black shoes, you know, available to everybody, you are going to know to go to sell him the same thing.
Mike
Okay, now you were touching upon that fit for sovereign local AI. Why is that better for those black shoes? Why is that better? Than a frontier model said differently. Why would an organization not just like, oh, it’s so hard setting up on my own local AI? I just give it all to Anthropic and OpenAI. Why would sovereign local open AI be interesting for that enterprise?
Radu
Because otherwise the A once the AI knows about that black shoes pattern, everybody in the world is going to know about that pattern.
Mike
Ah, okay. So what you’re really talking about.
Radu
And then I can come with no money, you know, just my hands in my pockets and open, you know, clothes and say, to whom should I sell black shoes? Oh, to Mr. John.
Mike
John.
Radu
John, here’s my black shoes.
Mike
And just to signpost this. And to make sure I am understanding, what you’re talking about is if I let a frontier model process my enterprise data, they get access to my data, they get access to my proprietary advantage. So I am by going to them losing that. And what you’re proposing is bring it in-house, get it back on the bare metal, run your local model. Govern, understand, control that data, protect that data, that’s your advantage.
Radu
Now that comes with a huge cost.
Mike
Okay, tell me about that.
Radu
Now that that comes with a huge it’s so much easier to to go and ask Claude and say please come in, scrap my data. I’ll give you two hundred bucks and tell me what to do. I don’t care if you are gonna tell two thousand other people. And that’s the startup way of thinking, that the startup mindset. And startups are doing this. Obviously, they don’t have data, so they don’t care sharing. Nothing. So, yeah, absolutely. There comes a point where you are, and we’re speaking enterprise level now. It comes a point when you are an enterprise. You cannot afford to do this. You are going to lose any battle in the future if you actually volunteer your data to the world.
Mike
But this is very exciting because what an enterprise can do is keep a full market advantage, historical advantage that was there way before AI, way before large language models, and continue to protect and nurture and to feed that advantage. So this is. This is really interesting because what it does is it starts to also signal that for the enterprise, they don’t just have two options now.
Radu
I think the two options that they have, the population.
Mike
I’m sorry, I mean OpenAI and Anthropic. It’s the choice is now far beyond that.
Radu
Oh, yeah, absolutely. What I want to double stress here is when we say internal, it’s not on-prem, physically on-prem. I mean, it can be anywhere in the cloud, but it has to be, we have to make sure that my data is my data and is not being used to train, you know, all the ways.
Mike
And doesn’t this control really help push forward the agent, the AI agenda in the enterprise? Because. Governance, assurance, control is in the hands of the business and not in the hands of a third party?
Radu
It has eventually it has to become like this. We’re not a one enterprise, the full globe being one enterprise with one mind
Mike
Well, some have proposed that they may be the only company in the future. I don’t know if you’re familiar with this.
Radu
Yes, their country as a company and the globe as a company. But yeah, we may not be around when this happens.
Mike
I think you might be right. But this is really exciting. So what I want to do is make sure that we grab this moment. So if we look at that full stack, the data is the proprietary advantage. And what you’re painting for us is a picture where you can fully deploy your local open, your sovereign AI solution around that to protect, to sustain and even grow your advantage in the market.
Radu
And that solution being also very niche on your vertical.
Mike
Yeah.
Radu
Yeah. Which is, I think, it’s very important as part of the process because it becomes more efficient, way more efficient. I mean, it’s. Tens of thousands of magnitude order more efficient if it’s focused. And yeah.
What to deploy now
Mike
So just to kind of frame what we’ve already figured out together, I think we’re definitely, if we take a step back, we’re definitely in another big hype cycle. You know, we’ve seen the cloud. Cycle. We’ve seen mobile, we’ve seen the internet itself. And obviously, before that, we saw the PC was a whole cycle on itself. The microchip was a cycle before that. So we’ve seen all of these cycles. We know that there’s absolute conviction that. AI is a thing, it ain’t going away. But what’s really interesting is the clue that you’re giving us is that really the answer to the future really lays in the data, the proprietary technology, and harnessing around that to particularly for the enterprise in your vertical. Your control, that’s like the sweet spot for future success.
Radu
I do think so. Now, that’s in the future.
Mike
Yeah.
Radu
I think that you cannot deploy now. Now, the question is: what do you deploy now if you are an enterprise, right? Because there is no way. And I mean, look, it’s very nice that we’re sitting here together, you know, a very nice evening in Bucharest, you know, drinking a little bit of water and fantasizing about future and philosophy and so on. Reality is that if our friend, the bank, needs to do something tomorrow about AI, and they do need to do something tomorrow about AI because data is going is flying from their things, and then they do what can they do? What do they do? Because they cannot build everything that we’ve said it now. There’s no tools. And then. We’re back to implementing tools that can handle the existing models today, which are the Anthropic and the OpenAI and whatever, but handle them in a way that they don’t that A they don’t kill them. B, they help them. And C, they prepare them with the technology to be able to build what we’ve discussed in the future.
Mike
And what I imagine the picture I had in my mind as you were describing this tool, working with all of these different agentic models, agents, tools, etc. What also, particularly when we go into the enterprise, into banking, insurance, and logistics, I mean, the other thing that’s on the almost on the other side of that line is all the legacy systems.
Radu
Oh, yeah. Have all the COBOL, all the all the. You know, yeah, absolutely. And you cannot afford to take them out. I mean, you are never going to take them out.
Mike
Why should you? Why would you?
Radu
No, no.
Mike
Particularly when you realize that you can, in fact, find the tools and the thinking that can connect those very agentic, forward-facing tools. And put them and get the plumbing into the legacy system.
Radu
Absolutely. And that’s the the kind of tools I’m speaking about. Because it’s a tool that needs to A, connect the two sides, but B, connect them so well, so it provides, you know, limited access and on on time and scope, and then one single point of connection and auditability and make sure who does what and all the all the good stuff.
Why FlowX.AI, and how big it could be
Mike
Well, we’re kind of matching. Right forward towards this topic of what you’ve seen in FlowX.AI. And just for some context, before we talk about FlowX.AI, with all of your different ventures, Either directly or indirectly, how many startups do you think you’ve been involved as an advisor or an investor in?
Radu
Oh, so so I’ve started as a founder of thirty-seven companies, six of them hugely successfully acquired, ten of them didn’t lose all the money and all the others amazing learning experiences.
Mike
Yes, the learning bucket.
Radu
Absolutely right. Then if you want to add to this companies I invested in and companies I advised, I think we’re getting above 150.
Mike
Okay.
Radu
Companies that I know people by the name, and you know, I’ve been involved in actual advising.
Mike
Okay. So that sounds to me like somebody who has to work across a lot of different companies. So if you’re going to add one, I think that’s a big ask because you have limited attention. And so I’m very interested to know what did you see in FlowX.AI that got them onto the list? Because I imagine that every day you’re saying no to many of those. What was it that got the yes?
Radu
Multiple things. And there have to be multiple. I mean, it’s not enough, only one, right? They’re in the right spot with the technology, with everything that we’re speaking until now. They are right in the middle of where the infra, where I believe. It may be wrong, but where I believe that the infrastructure for AI for enterprises should be, right? So they’re really like fighting head to head with Palantir, with Stripe, with Okta, and so on, right? And that is amazing. Second is that. I found in Ioan an amazing founder, the right type of a founder, relentless, energetic, articulate, with a vision, with the ability to execute business, with a demonstrated ability to execute business. You know, to be honest, there are not so many founders like this. And that’s interesting. So the the right technology, the right vision, the right founder and the right backing and support VCs and
Mike
The community of stakeholders behind that.
Radu
Exactly. Way more articulate than me at this time. So yes, exactly. Because it’s not enough to have technology, it’s not enough to have the founder, it’s not enough to have the team. You need to have all the stakeholders that has to be there to help the company.
Mike
I mean, it reminds me the idea it takes a village to raise a startup.
Radu
Yeah, exactly, exactly. So I’ve seen all this in FlowX.AI, and I’m like, and look, I’m as egoistic as every other person. I’d love, I’d prefer to be connected to successful companies rather than. To learning experiences, right?
Mike
That bus is full.
Radu
Exactly, exactly. So yeah. It and it sounds like a it sounds like a winner. Happy to come with my two cents and help to this winning proposition.
Mike
Okay, so my last question for you, Radu, is what do you think the real opportunity is for FlowX.AI to help its enterprise clients? If it could solve a really big challenging problem for customers, what do you think that’s going to be?
Radu
An ideal world? I’d love to be [unclear] and go to the, you know, [unclear] or CISO or whoever in any bank and say here’s what we have and with this all your people are going to be able to use all the AI in the world. Your data is not going to go away. You’re going to be able to know and to manage and to approve everything happening to with AI the same way you are managing, approving everything that’s happening with the rest of your data today.
Mike
So that seems like it goes from being not only a tool, but it gets close to almost being an operating system. Yeah. I mean, we’re dreaming.
Radu
Yeah. Yeah, depends on the definition of our operating system, but yeah, it has to be the actual connecting layer, the only way, the entry point, the exit point, the audit point, everything that relates to AI
Mike
And how big do you think the impact would be if it realizes a dream this big, if it is truly the integrated view of doing business within the enterprise? Yeah, how big is the impact here? I mean, what do you see in that future?
Radu
Man, look, we’ve seen companies. Solving lesser problems growing from one dollar to ten billion dollars in two years and those were lesser problems than this. Is this a ten billion dollar potential company? Yeah, it is. It is, and it’s more. It’s now about actual execution. So there are all the seeds. You can see them happening. There’s a long way going forward. Hopefully it happens.
Mike
Well, what you’ve actually painted for us is fascinating because we know that. There is this cycle, this wave that we’re catching. Could get a little bumpy, but beyond that is a huge idea to become the tool of choice within the vertical with an open, sovereign architecture and that deep vertical expertise for each and every client, allowing them to be that true middle orchestration and integration role. Spot on. And allowing not only Ioan, but all of the team here to access that community of stakeholders, clients, partners all around the world. And the impact could be big.
Radu
Could be huge.
Mike
Well, that’s a great place for us. To wrap up today’s conversation, Radu, thank you so much for coming and having this conversation. I hope we can have another one in three months from now.
Radu
Absolutely. Absolutely. Thanks so much for having me. It’s amazing. Always enjoyed fantasizing with you. And yeah, look forward to that.
Mike
Thank you so much. Thanks, Radu. Thanks.