 Welcome back everyone to the live state coverage here of SuperCloud 4 in Palo Alto. I'm John Furrier, Dave Vellante. I've got a great day going on, great content, two days tomorrow all day streaming as well. We're unpacking generative AI and all the hot topics, technologies, business models, and of course startups in this panel. We've got two amazing companies, startups, young, hot companies. Here we go, here we go. Here we got Amit Yelkovin, CEO of Cubia, which is Cube, thank you very much. You'll get a shift in CPO of wand.ai. Gentlemen, I want to say thank you for coming in today and thanks for coming and contributing to our program. Thanks for having us. And I want to say hearts out to your fellow countrymen and women. I know it's a tough situation. We're here to talk startups, but we hope things go well. Appreciate it and we appreciate your support. I know you've been good friends of Israel and the US in general has been very supportive. On behalf of Israel, we appreciate it. Well, we're fans. We know a lot of startup activities. It's just a little bit disruption. We hope we get back to normal soon. Like 9-Eleven will never be the same, but business still goes on. You guys got awesome gen AI activity going on at Cubia. We've talked before, we had a great article on SiliconANGLE and enterprise AI is hot. You think like enterprise slow to consumer, but there's a lot of data in there. David, I've been covering. You guys take a minute to talk about your companies. Let's start with Juan first. Yeah, I'd be happy to start. Thank you for having us. One AI, what we do as an enterprise assistant, basically a business assistant, what we're trying to do is taking generative AI and enable it for enterprise companies. Basically, strike down any barrier for entry, for generative technologies, generative AI technologies for those kind of companies. Specifically speaking, world-based security and the ability to access all the organizational data, but only the data that is relevant for you. Second can be real-time access to data. The current technologies doesn't allow you to operate with the scale that requires to operate in those enterprise environments. We developed a unique search capabilities that allow you to interact with large scale of data. And we also added a personalization framework on top of our assistant that allows it to adapt, basically to the interaction you're having with it and allow you to tell it once what you want to do and it will understand your day-to-day actions and also execute them. It allows you to send emails, set calendars, run analytical tasks, predictive AI, whatever you need to perform your day-to-day business and also your analytical business. We'll get into some of the hard problems that are now being solved. And the price has always been hard with search and data, all the silos. But when did you guys start? How big are you guys? What's the status? Where are you guys in terms of your progression? The company started a bit over a year ago. We had a few pivots around the way. Finally integrated predictive AI, and generative AI together into one product. We're almost over 20 employees, most of them in the Bay Area, same Europe and in Israel. And we are currently expanding and having a lot of conversation with big enterprise companies, actually getting the product to the end line, trying to sell to the business user. I mean, I talked about your company, I know we have it on theCUBE before and also SiliconANGLE, but for refresh, you're a platform engineer, you can be a KubeCon, we'll be there as well. You got an interesting platform engineering angle at Jenner Bay Eye. Absolutely, I'll do a little quick recap for those who haven't seen my previous interactions on theCUBE. We're looking to solve a very specific exchange between the developers and the platform engineers slash DevOps. The concept is when it comes to self-service for organizations, how to unlock the golden path, how to essentially serve yourself to infrastructure, automations, knowledge and so forth. That's not always intuitive, that's not always as as straightforward as it may appear. Imagine that even in 2023 with large language models, there's still a lot of adoption barriers and we're seeing that plateau. Most organizations who pick up internal developer platform or try to develop themselves usually have a 10% adoption barrier. That's typically what we're seeing even with backstage and technologies out of nature and our thesis all along has been how do you go and unlock that? How do you go essentially democratize the usage of DevOps or technical functions in organization in a way that's intuitive for folks? And the common denominator in all of this is natural language, right? Large language models unlock this, but how do you rein in large language models in a way that's predictable outcome? Our users are very unforgiving. Especially for mistakes. You don't want to have to wake up one day that your production is down because a large language model had a hallucination. So this is kind of where we come in with our own co-pilot approach where every user has their own co-pilot, not a code completion co-pilot and an operation completion co-pilot and we can talk about that. I love where you guys are right now and I want to just real quickly get a status. You guys had raised some pre-seed funding? Yes. Okay and your funding level was series? We also had some pre-seed funding, safe funding. Same. Yeah, so this is, we had a panel on earlier on the founders with more of the, Sure. Senior guys. That's who I was telling you. The old guy founders and now we get the kids. But this is why I like what you guys, you both mentioned the word pivot. You mentioned it to me privately, said pivot. You guys started your companies right at the right time when this wave comes. Talk about the agility of the startup teams. From an opportunity recognition perspective, obviously you start a company because you have a vision. You see some opportunity, you want to capture it. And then you've got market conditions. You're in the arena, right? You've got to make moves. And then the AI wave comes in and you see an opportunity to take your, I mean your vision was Siri from DevOps, right? And then wait a minute, it's not Siri, it's AI. Right, tailwind for you. And enterprise AI complexity, the old enterprise success model was solve complexity with more complexity, right? So that's not what AI is doing. So talk about your pivots, how you saw it, the mindset. Yeah, I can address that. We started with predictive AI, creating predictive models for analytics for data scientists. It's allow enterprises to basically connect whatever they want to create a data science pipeline to get analytical results. What I just said sound very high level, but when you interact it with the human intuition generative AI allows you, you can get a human intuition on solving data science task and action. You can actually just ask us what is my turn and I will get the answer and not try and let someone build it for you. So that's the pivots we did. We saw the opportunity of automating everything we wanted the human to do because the human part of what we tried to do is not the complex part. The IP is in solving the analytical part and the human interaction is just another tool to allow it for the enterprise. So take that example of churn. There are many ways actually to calculate churn. You can do custom, you can do NRR, et cetera. So how do you ensure that it's consistent? Just as an example, I mean, I'm sure we could use many examples, but churn is a good one. Does the customer sort of train that or do you sort of have pre-trained models? The real answer is what Amit said is generative. I also opened the world to having mistakes. Okay, it means that I will give you a churn based on your data and the way I think you can solve churn with the data you provided me with. But if you're missing data, you'll get maybe the wrong answer. Maybe you get a slightly wrong prediction, but we're working really hard to not get a hallucination. That's the difference between getting the wrong prediction that even data scientists can get and having the wrong elucidated prediction. I'm chuckling because garbage in, garbage out, when people connect their docs. For example, to us, we're like, why am I getting this answer from a document from three years ago? Listen, I'm not writing your docs. I'm giving you the retrieval capabilities for it. Talk about your pivot too. There's a moment in time where everyone has that moment where it's like, okay, the bit has been flipped. You go and you got to get your teams. Do you turn around to your teams say, look, we go here, we're going to go faster and the trend is your friend, as Dave always says. In this case, it's legit next level AI is here. What was the moment where you saw the AI, what made you cross over? So from you correctly pointed out, we discussed the concept of Siri for DevOps and people. Siri for DevOps, what is that? Think Siri just for DevOps, like what does that mean? Conversational AI. Provision that Kubernetes cluster. Yeah, right. And then- And this was before chat GPT. Correct, about 45 days before and 12 hours. You remember the exact moment, okay. And then chat GPT came to a world. And then the biggest barrier of entry for us, the market education, we didn't have to concentrate about that anymore. A million people in 48 hours all of a sudden we're on the platform enjoying conversational AI with the machine. Now, where do you take this? We already had all the scars of training our own models, of effectively creating rule-based systems, merging them with LLMs. How do you go and now reign in on the business potential of this? And we saw the agent-based framework as the way to do it. And this is essentially kind of the concepts here. People, even though they enjoy talking to intelligent machines, they still want a bi-directional interaction. Workflow-based systems are one directional. So how do you go and create the near-specific thing to a human interaction with the agent framework? Obviously, you can start a chain of reason, trivial augmentation. You can go through all the different practices here. Then a day, it's a human interaction. Human-like interaction. And your firm, your employees, same as you'll give as you're in Israel, you're in... Ironically, we were almost twins from different industries founded around the same time, around the same funding, around the same people. And we actually grew up not too far from each other without realizing it. So... I'm glad to have you all here on theCUBE. And so you did your safe. Did you do your safe in Israel, or did you do it in the States? Both, for most, but yeah. Okay, both. So could you talk a little bit about how you did that and what your thinking was? The first Israeli entrepreneur I met was the very famous Moshe. You might have heard of Moshe and I. And so then I got interested in Israeli companies and I'm like, wow, these guys are pretty smart. They own Cyberspace. But you've seen both models. And increasingly, you want sort of a global presence and you certainly want a presence in the Bay Area if you get it at Silicon Valley. So how did you go about that? Maybe you could tell your story there. About how we got to the... Yeah, your decision and how you thought about the funding and how you recruited different investors from different parts of the world. Yeah, I think it also relates to the previous question. I think generative AI and AI research in general are focused in the Bay Area. There's like between Stanford and Berkeley, there's a lot of smart people that are researching AI specifically. And I think that was what caused us to search people around here from the talent perspective and with that, the investor's perspective. The Israeli front came from personal connections and history we had at previous companies before. I came from the Israeli military. So my connections are deeply inherited in Israel but also in the Bay Area because that's where the high tech communities goes to. Like there's a rich community here. And your funding was the first round or the seed round, safe round, before ChatGPT or... Yes, yes. It was, okay. So wow, that's your right. Well, that's perfect. Well, when was your time that you had the pivot? Because there's a go-to-market change you mentioned because ChatGPT educates the world and what you're trying to educate people on. And then is there a product change? So imagine there's now product tailwind as well. Yeah, so I think on our end, it's changed us from selling that... The change came from the client because of ChatGPT. The change was, for us, it was instead of having another product for data sentence, I can actually sell to business persons. I can sell to the sale persons. I can sell to the customer support. I can actually create value to the end user which was the missing part of being another tool in the chain of tools that people are using. And the middle person, the data scientist, the go-between, between the ultimate end user. So the democratization or consumerization of AI that's happening is really what's happening here, isn't it? So you had increased TAM and you had education overnight for your consumers. And a lot of noise. Yeah, but so that's why I wanted to ask you, from a fundraising standpoint, in the one hand you say, wow, now these guys are AI, there's a great tailwind. It must be really easy to get funding, but there's now it's so noisy, so crowded, so much competition for attention. So what have you experienced? So how do you differentiate what you do from somebody with a thin layer, API call, simple API call to chat GPT? That devils in the details. You mean AI wrapper app? I'll call it AI. I'll call it AI. Vaporware, whatever you want to call it. Well, I mean AI wrapper, it's like a website. I mean, it's not bad if it's a use case, but it's not a scalable, durable company. Not a feature. There's simple Lipton tests you can put to just have a conversation with it. See how fast it breaks. See how fast the interaction goes into areas you probably wouldn't expect to. Scale. Scale is one aspect that's more in the infrastructure side. But quality. Quality, interaction, context, remaining context, having a state machine that knows how to keep the context and orchestrate the business logic and APIs and the workflow less workflows in this. That's different. There are different levels to this. I don't want without having to go into the details. A lot of people are now discovering all of the challenges of trying to tame and... How do you feel about your current solution, VisaVee, as you look at what you guys have built VisaVee, some of the vapor, well, I'll say vaporware, but it's like demos. Because right now there's, people are trying to squint through the demo problem. Yes. Everyone, hey, it looks the same to me, but like you said, what's the test? How do investors know that you guys have a better deal than someone else or a customer? I tell you what we did. We just released it as a self-service. We said, you want to try it out? Try it out. 700 people signed up in a month. And we understood that there's the positive feedback here. Yes. We're building. We're always building, right? It's always going to improve every day. You're on the platform as you said, which is a little bit of bottoms up kind of strategy. You've got the enterprise, so you've got customers telling you what they want, right? We actually got both of them. We have a big community on LinkedIn. We have a few of the largest influencers and groups on LinkedIn, on generative AI, that we're working with. So we also have a B2C product that is used to show people, like Amit said, the value of using our technology. And then they can imagine how it will be with organizational data. They can connect their own data. So synthetic data? No, they can use it as a personal assistant and then see the value and then say, OK, I will use it in my organization. OK, I got it. So a freemium model. And the data sources for the enterprise customers that you're talking about come from the enterprise customers, yes? Is it all happening in the cloud? Are you getting customers to say, we want to do this on-prem because we're afraid of IP leakage or privacy? It's more of the latter. Companies want, especially the big enterprises, they have a huge conflict. They've been told to invest in generative AI. It's an advanced technology. But they are, as you said, still like a... And they're set up to underwrite it. That's un-funded. It's unfunded so they have to steal from other areas, right? Or not necessarily? No, actually it is funded. There's a lot of companies who just bought racks on racks of GPUs to use for something. And they don't have any application to actually operate it. So we do provide... We need to ask. Help out. Process-wise, right? Organizations still don't know how to underwrite the concept of GenAI. They'll come and say, we want to go and use GenAI tool. And then their security team will block it out at the go. And then we're like, are you using GitHub Copilot? Yes, we are. Well, what gives? OK, I'm a VC. I got a big bag of money. I want to write a check. I want to do your series A round. I got a couple of questions for you. Sure. What is your vision? What do you guys vision? What's your North Star? What's the big idea? What's the big market you're going after? What's the big vision? For our perspective, we feel that we're really in the crossing point of where data, operations, and language start intersecting. And it's only a matter of time until we'll have enough data and enough essentially feedback loops with our customers to create the first self-driving DevOps, essentially digital DevOps, digital employees. And that's one flag. We can go and plant that and say that goes into other areas than the technical operations. But it's a proof point. And we're seeing where this is going to go. Your vision for your company? Our vision is similar on the business side. People nowadays consume it back by researchers. It consumes almost three hours a day just looking for information, tagging between different apps. Just look how many apps you're going through on your phone or on your laptop without actually doing any productive action. We think that the generative can really disrupt, for your question earlier, disrupt the way to productive actions on data. The huge amount of data people collected, the big data era, and a lot of data out there can now be consumed. And we want to be the one to facilitate the productive assistance for you for any task you want to perform on your business day-to-day action. And personalization must be really important, obviously. Yeah. A big part of it is personalization. Really taking your data and connect it to who you are, who you are in the organization, what is your contribution to the organization, or how do you like to interact with it? Nowadays, when you interact with a bot, we call it passive personalization. You tell it what to do, and eventually it does whatever it does. We develop what we call active personalization. We can actually learn what you're trying to do. And next time you're talking to me, or even if you're not talking to me, I'm with through your emails myself, and I understand what you're looking for. I'm your real assistant, like your real EA trying to give you insights of what you do. I will say, like I said before, I will be your real EA. I can be good in some stuff, but I can give you not what you expected on different stuff, but I will give you the right. Hey, we got a term sheet ready to go. Jump full, who's going to get the term sheet? Well, the next question is, who's got traction? What's the traction? What's adoption like? What's the customer, yeah. Our new co-pilots have really taken off. We've seen that people are enjoying that natural interaction. Native querying, native actions without needing to set up anything in advance, any workflows, any gated. Only until that point, then they want to go to that next level, which is how do I go and reign it in with governance, with policies, and so forth. You have good traction. Any momentum numbers, can you share? What's some of the? Probably let that to the imagination of the tier one VCs. I'll ask it. But you have people using the product? Yes, yes. We have hundreds of people in the system. Yeah, as we said, we are a freemium model. We already have over 10,000 users in our platform. A few hundreds of them are already pain. We started monetization only a few months ago. And we see a lot of traction. And it's a very dynamic world. That's, I think, the biggest pain point. Like Amit said, people come in, and they don't know what to expect. And you need to keep them in. Our internal metric for ourselves is how long until someone is living with the platform. I told everyone in the company, if you can't stay in the chat and have a productive chat for 15, 30 minutes, then we failed. We need to figure out how do we get there. So that's your North Star metric. How long they stay engaged? And your productivity. And you're seeing that type of, well, I'm sure it's a bell curve, right, but you're seeing that type of adoption. Yeah, we see. And those are actually super users, if you like. Give us a lot of insights, because we see actually how they interact. We talk with them. We create, we do client interviews to see what they do and how they interact with the platform and what they actually want to solve with that. Because the use cases are infinite. Because it's a human interaction. In any front, in the DevOps front or in the business front, you can do whatever you like. People sometimes guide us, guide us what you want to do. It will do whatever you tell it to do. The limitation is only your imagination and what you're trying to do with it. What do I get for free? How do I get my data in there? What's that adoption for the customer look like? We limit it right now in a few angles. You can limit it by usage if you like to talk with it a lot. We limit it by the number of data connections you want to put in, because storage-wise it takes a lot out of us. So you have different plans, depends on the usage. But for 30 or 100 bucks, you get different packages that allow you to chat with it forever. How do you integrate into things like Salesforce or Slack? I know you guys have Slack integration. What's the integration like on your end? For outdoor integration, we also have Slack integration. You can also embed it in your website. We have a lot of clients who just scrape their entire website. You can scrape websites with the platform and put it just as a chat bot on the website, because you can create bots with one. One of the features that we do, we call it 1D, that's the bots you're creating. You can create a composition of different connectors. If you are sales 1D, you can connect your sales force, you can connect Zoom info and connect your email, and it will create client calls. And we get you prepped for client calls before you want to get ready. Dave, what's interesting is you got kind of the business productivity, tech integration there, which is kind of tech stack meets business integration. And then you got platform engineering, a little bit different market. I see the numbers probably higher there, but it's the platform engineering that DevOps crowd. They're a tough crowd because, well, they're not, they're a good crowd. They're great people who go to cover the open source, but there's a lot of nuance to the infrastructures of every single company. So in a way, your adoption is a little bit harder, I'd imagine, because they got to get comfortable with the tool and make their environment connect to you. It's how do you- There's higher stakes, typically. Yeah, take us through how you land the beach head, because your beach head is clear, problem solved, clearer to me at least, but not so clear on the DevOps side, because I see it being squishy there. What's your- Actually, the co-pilots really help put all the puzzle pieces together. So the co-pilot approach is, how do you go and interact with the bot in a way? And we can give you examples. And AWS co-pilot approach would be, create an ECR, attach a policy to it, doing all sorts of investigation and context switching around AWS. And then you want to go and list all your GitHub repositories. And then you want to go into JIRA and interrogate your JIRA Ticket queue. Now you want to go, I don't know, into Kubernetes and create a deployment or restart a deployment or view logs. So- So there's productivity angle there big time. Absolutely. And that's usually the first flavor, the pellet that gets wet. And now you want to go and put some guardrails around that and you want to have more templatized infrastructure for an automation for the rest of the organization. And that's where the internal developer platform puts on top of that. What's been some of the feedback? Can you share some of the anecdotal comments from your customers that, because this is really getting into the Kubernetes containers, the DevOps, DevSecOps, I can imagine, that's vectoring into security area. Yeah. People are amazed that you could have these type of interactions. I'm not your most technical CEO, but I can go and create deployments myself and natural language interaction. I think that's a wow effect that people will have been- It's a good party. Hey, spin me up a Kubernetes cluster. Yeah. What's Kubernetes? Now it's also scary for some people because they're saying, now there's so much power with these tools but how do you rein in, that's where the governance comes in. I think the governance compliance we've been seeing, this has been the big trend in SuperCloud 4, is that if you get governance compliance right in the front end, everything scales, the intellect scales, the automation scales, the productivity of opportunity scales, that seems to be the quick main thing that everyone has got to do their homework on. But a lot of times people just, they're out experimenting and they think, okay, they bolt on governance and privacy at the end, but it seems a little bit different this time because the lawyers are more involved. Companies are more concerned. You've heard some stories that oops, we should have done that. So I think there's a greater awareness. Would you agree? Yes. And as you have also mentioned, so we have a hybrid infrastructure where people can run the entire operation execution side of it from their infrastructure. And then they call back our large language models and our embeddings. All right, final word as we wrap up. First of all, thank you for coming on to the Cube for SuperCloud 4. I love the hot startups. You guys are in the, I'm kind of jealous. I kind of have fun while I wish I was back on the trenches, your age starting a company, because it's a great time to start a company. Advice to other founders out there and around what's going on in the arena as an entrepreneur. I mean, it's very challenging, very difficult to do a startup. So hats off to you guys. And two, what do you hope to achieve this year? So two questions to end. I mean, we'll start with you. Advice to other founders how to handle the current situation and what your goals are in the arena. Stay the course. It's we're in choppy waters. I don't even have to go into the Israel situation. You addressed it up front. Certain things just belittle everything else we're doing, but it's still stay the course. Markets are gonna go up and down. You can't control what the VCs, what their appetite is, what the flavor of the month is, what their evaluation is. You just have to go into execute and control what you can. And that's what we've been doing. Keep your boat away from the rocks, as we say. You know, where do I want to go? My next milestone and the next 100 customers. And then beyond. There's a lot of smart money out there. A lot of smart money. So, you know, even though the funny market may look weird, it's weird, there's some smart money out there. But your advice to other founders who are in the arena and what's your goals for the year? My advice is again, execution. I think that's the one thing startups needs to have in mind, not investors, not like the markets. You just need to focus on your execution. Because as you said, we are agile. We can move fast. We can break stuff. Focus on your execution and differentiators. Like the side of the value that you bring to the market that no one else does, even if it's small and make it grow. But focus on execution. Because if you can do it good and fast enough on that differentiator, on that very small differentiator, you will win. And you will get the money you need and you'll get the people you need to do it. And your goals? My goals as any business is to be as big as we can and hopefully not needing any of the investors money. We hope to grow fast enough to outgrow that but also look forward for a partnership with investors and clients. It's a great market. It's a whole nother. Cloud was easy to get started and now it's easier with AI. You can, if you get it right, that's the key. Guys, thanks for coming on. I really appreciate it. Great to have you, Q. We'll be back next with our enterprise guru panel. Howie Shue is going to be moderating an excellent panel between Google, Microsoft and Salesforce. Top AI leaders to give you the perspective of innovation at the enterprise. We've got the founders, we've got hot startups and now enterprise leaders. We'll be right back with Howie Shue and the enterprise leaders panel after this short break.