Elias Torres wants to reach unicorn status with just 100 employees

0:00Starting a business comes with its share of ups and downs, which is why staying true to your vision is essential. a non-negotiable for Romeo and Milka Bregali, Capital One business customers and co-owners of Ra's plant-based restaurant in New York. Romeo and Milka took a leap of faith when starting their own restaurant, gutting an empty space and building it from the ground up, every pipe, every wall, every detail. But building from scratch came with a heavy financial burden, which is when they turned to their Capital One business card. With the flexibility of the card's no preset spending limit, they were able to spend more and earn more rewards while bringing their vision to life.

0:36Today, Raz's success is proof that with passion and the right support, it's possible to make your dreams a reality. Learn more at CapitalOne.com slash business cards.

0:53We're at a peak moment in the history of the world of society, especially around business. In what way? This is the most humans will ever have at companies. Servicing the customer. There will never be more from this day on. Elias Torres is the kind of person who speaks in definitives. He's direct, innovative, and future thinking. These qualities make him a successful founder. What I find especially interesting is how he's building his company. Elias is trying to reach unicorn status with only 100 employees. It's a lofty goal, but considering his track record, it seems possible. He's a Hopspot veteran and sold one of his company's drift to private equity for$1.2 billion.

1:44But our conversation covered so much more than just business. We talked about Elias' immigrant story and how that experience shaped his career. Maybe because I'm an immigrant, maybe because I always took care of my family and my siblings and my mother. Like, I'll take care of this problem. I'm a wartime CEO. We covered the American dream, and if it's still alive. I came to the United States trying to conquer the United States. I did it. And now I'm like, the food here is poisonous. We're not healthy, too stressful, too much work. And I go back to Nicaragua and it's a paradise. And we also explored how AI is changing our ideas of what a fulfilling career looks like.

2:30I think that's what it means to be human, right? Is to be authentic. To be, I'm doing this because I want to do this. I'm Rana El-Khalyubi, and this is Pioneers of AI. A podcast taking you behind the scenes of the AI revolution.

2:57hi alias thank you for joining me today i'm so excited for our conversation absolutely thank you for having me it's gonna be fun let's make it fun so before we dive into ai i actually want to start with the human um this has kind of been my journey like human-centered ai and i want to start with one particular human your mom okay you have a podcast on the side the American Dream, and I was listening to your interview with your mom for Mother's Day, and I just thought it was really heartwarming. My mom has played a huge role in my life as well. We're both immigrants to the United States, so I wanted to start there.

3:33Tell us more about your story, how you got here. I got here because of my grandmother, my mother's mother. She came to the United States, crossed the river, El Rio Grande. So like the true, true, like she just, she was 65 years old, got on a bus, walked. I have no idea what she did in 1975. And she was, she got to, and she got to family in San Francisco. And she, from there, as an illegal immigrant, found her ways to learn about community and helping the Latino community became some sort of activist. I don't know, but she ended up getting her green card. Asylum did this, you know, later in her years.

4:23And because of her, she was able to ask for my mother and my siblings to come to the United States with a green card, right? And so we waited like 10 years from 1983 to 93. That was the process. and I can't believe that happened. This is no internet, nothing. And one day, my mother goes like, we're moving to the United States. And I was just so, so excited. The American dream was like just, I didn't even know it was called the American dream, but it was like in me, like, mom, we're going to go. We're going to work. We're going to do this together. We're going to build a home. I was so fired up, 17 years old.

5:07and first thing is we did it was like I always tell the story I cleaned offices I was vacuuming I was coming at night with my mother but that was my story that's I always say that and then my mom on the podcast goes you only did it once okay I'm just like I mean I can't edit it now this just ruins me it's out there now so but I did it I did it once worked at McDonald's worked at a grocery store, Albertsons. And then your dad somehow got you access to an IBM computer. That was early on. There was a seventh grade time when I was in LA. I visited and he got this computer. He, you know, he would buy stolen stuff.

5:47We don't know where these things come from. We don't ask questions. Just don't, don't worry about it. But I had no idea what it was. You know, My only connection to it, it looks like a typewriter. That became a little bit my obsession. I don't know why. I just loved it. Even though I didn't do much with it, but I just kind of like fell in love with it. I just love the idea of doing things more efficiently. So that exposure to this early IBM computer kind of set you on this journey to fall in love with technology and pursue a career in tech. And I always like we're both parents and I always think about my role as like a chief possibility officer or chief exposure officer for my kids.

6:29How do you make sure or ensure that your kids are kind of on the forefront of of AI and technology in general, but AI specifically? My kids are are older now. Right. So I just graduated my youngest out of high school. And so he's going to do CS. Actually, all three of my kids are CS. How did you do that? I am not able to get my kids, maybe my son. There's hope. My daughter's not into tech. All three did it. I don't know what they're going to do, right? Everything is changing. I don't know what their real passions is. But I'm glad. I think it goes back to your question, right? Of like, how do we expose them to things?

7:07And I'm biased, right? I think having CS under your belt, it's like the ultimate superpower, right? It's like, it just teaches you that we can do anything. You can use technology as a tool to do anything. We can build anything. And so it's like, especially now, you know, in a distributed online world, and especially with LLMs, it's the ultimate superpower. I mean, I know everybody's excited because they can go to Lovable, they can go to Replit, and like, I can build an app. But really understanding how it works and just understanding that you're responsible for building it from, and you can, if it breaks, you can go and fix it, right?

7:43and just demystify that, but have enough ability to do that, I think is the most. So I'm really happy and excited that they're doing that, but they might end up choosing different paths, different combinations, and I'm encouraging. That's kind of been my shift now to be like, my daughter makes CS with design. Awesome. My second son is pure CS, probably AI, obviously right in. And then my youngest is the most creative. He's a painter. He's a drawer. Oh, cool. He's sports. He likes physics, kind of. So I'm encouraging him to be like, mix it with whatever else. Yeah, I like this idea of combining kind of a computer science degree with something that you're really passionate about.

8:27Okay, so before we move on to kind of your entrepreneurial journey, do you think the American dream is dead? Today is alive, right? I think it's alive. I think we're going to need it more and more. I can't think negatively, but this is what's going to happen to the United States. You know, I don't. I have to. I am worried in general about society as a whole. I think people are just tired, right? I think people want answers. People want solutions. I think what we need is more like people want to fix things. Yeah, hope and solutions. And there is a lot of hope, I think, that there should be. But I think the dream is a life.

9:07Our culture is of immigrants, of entrepreneurs, of rule breakers, of we take high agency people that have ambition and aspiration and they don't take no for an answer. So that creates room. Might not be equal, but there's still more room here than there is anywhere in the world. When we come back, we'll dive into Elias' self-made entrepreneurial journey and learn why he devoted his career to improving customer experience. Stay with us.

9:57If you've spent any time building AI products or leading technical teams, you know this. Transformation doesn't fail because of ideas. It fails because teams can't move together. Enter Atlassian's Teamwork Collection. It has planning in JIRA, documentation in Confluence, video updates in Loom, and now AI agents in Rovo, which connects the dots across your work so nothing gets lost. It's one AI-powered teamwork platform designed for how modern teams actually build. Learn more at Atlassian.com slash TeamChanger. That's A-T-L-A-S-S-I-A-N dot com slash TeamChanger.

10:47Elias worked at IBM for years as a software engineer, but he wanted to do more. He first co-founded a company called Performable, which he sold to HubSpot. He then eventually co-founded Drift, which helps companies better service customers through more direct and engaging conversations. Mind you, he founded this company before the Gen.AI explosion. And after nearly a decade, in 2021, he sold the company to Vista Equity for$1.2 billion. I wanted to know why he decided to sell and why, in spite of such a successful exit, he still wanted to start a new company. Well, 2008, I quit IBM to join David Kanzel.

11:33I want to be in the startup world because after 10 years at IBM, I'm like tied up in there. I'm going crazy. It doesn't work for me. Like nothing's getting done. Nothing's getting done. Nobody cares. I think that that to me is the most apathy. It just kills me. I only want to be with people that care. And so I go into the startup world and it's like working with David and it's like intensity all the way. So we did a company called Lookery. Then we'd start another company called Performable. We sell that to HubSpot. Three years at HubSpot, we take HubSpot public, a billion dollar company. And once you do that, it just felt too easy.

12:08So we're like, we can do this again. But that's the beauty of it. It's like everything in retrospect, one of my biggest advice that I would give to people would be like, nothing is as hard as you think it is. I'm not saying it's easy, but it's not as hard as you think. It's like, I think when people just like block themselves, they become their own obstacle. Yeah, absolutely. And so like, you know, and I'm not going to discount how much luck has to do in my life, but we did that. And then we came out and we said, how hard could it be if we did this with HubSpot? we grew from$35 million in revenue to$135 million in three years.

12:49What can we do this again? And then you start breaking down, like, okay, how many customers to get to$100 million in revenue, depending on what they pay, right? It's like 83 ,000 MRR. Monthly recurring revenue, yeah. So then we're like, you know, how many$1 ,000 deals or how many$100 deals or how many$10 ,000 deals? You need eight$10 ,000 deals a month paying you. Doable. All this is doable, right? And so we just came out with that mentality, David and I, and it was really hard, right? I mean, it took us like two years of wandering, we call it, through the desert to figure out what would generate such a thing.

13:25And so that's the journey that I was adrift. I think we found something which was really AI, before AI, which is automating, you know, people have these websites and they have no one there to respond. You fill out a form, contact us, and nobody responds. And so we were like, what if an AI agent could like talk to you in real time, book a meeting, route you to the right sales rep, ask you the right questions. So we were doing that automated on the website. Way before LLM. Way before the LLM, right. We were using BERT. We were using all the different models before the LLM came fully together. And we made a lot of mistakes.

14:03And so I kind of like took an opportunity. when the offer came, I engaged and I did not hesitate. A lot of people were like telling me, Lias, look how much money you're going to make if you're$2 billion. And you could be that in a few months. But that was September 21, January 22, the tech world collapsed. And so everybody got left holding the bag. But we exited at that time. And private equity is amazing because they buy you for cash, you know, and you're not waiting for sellouts and IPOs and stuff like that. So I call it sometimes like my biggest failure because like everything else, I think I exceeded the expectations.

14:43But this, my expectations were higher than where we ended up. Even though you did very well financially, but like emotionally. Yeah, emotionally I felt like, you know, it's like, I think as I'm getting older, like I want some sort of legacy now. That drive to build a legacy eventually led to his current company, Agency.ai. Right after ChatGPT came out, Elias spoke with a mentor about what to do next. He was financially secure after selling his company, but his mentor thought Elias still had one more company in him. His new company offers a way for businesses to leverage AI tools to create a better customer experience.

15:25As for what that looks like on the ground, Elias says it's constantly evolving. There are so many different directions AI can take you, and it's not always easy to cut out the noise. I talk a big game on LinkedIn, but the reality is that it's hard. Why is it hard? Because there's so much competition? There's so much noise. It's a fog of war out there. People are trying to figure out what to do with AI. And there's a lot of stuff like, look, it can generate a video. And people are distracted by that. How do you drive real value? How do you drive real value? How do you really scale? And so like the idea is that you have to think about the full scope of the challenge of the problem.

16:07So to me, it's customers, right? It's the overall customer experience. That's what I'm solving for. Not just customer support. I think actually that's a really important distinction. Your agents aren't just coming in at the point where a customer needs to solve a problem. You're thinking about the entire customer experience. As long as there are businesses around, we're going to have to deliver experience. And that's what I want to solve. That's what I want to solve for. That's what I want to build. I have no boundaries. I have no rules. It's the customer experience. Well, I remember when we were together at the Fortune Brainstorm AI conference, you pulled up agency.

16:46And I guess agency is there to support you, even as a CEO of a company, to like track who are the customers you're talking to, what are their needs. Like it's not like that struck me in that conversation. It was like a one-stop shop for all things customer related. Is that kind of the vision? Yeah. You know, we have to build companies with less people now than we ever did. Right. I mean, like I'm living that. Right. I mean, like we can't. I made a promise and I said to Sequoia when we did the seed round that I was going to get to a billion in revenue with no more than 100 people. How many people are at the company today?

17:23Like 15. What's the breakdown of that team, by the way? Let's say 12 engineers and three non-engineers. And then a lot of AI, I assume. A lot of AI. Okay. I've never seen code written so fast ever in my life, right? And it's like, people are asking me, like, you know, yeah, if you have 10 engineers and without AI, they can do X. But if you have AI, those can do 10X. So do you go to one engineer because you can do the same? Do you keep the 10? Or do you hire another 10? Which one is it? Which one is it? Which path are you on? I want to go faster, right? So I'll hire another 10 and do twice as much as the other 10 with AI, right?

18:11And do more AI. And so you got to start things earlier and you have to try them out and you have to build more and you have to experiment in parallel multiple times and faster because the world is changing. And right now is the most chaotic moment with AI. If you can get there before, that might give you a huge advantage. After a short break, Elias and I talk about the personalities of AI agents and why agency's agent isn't nice. Stay with us.

18:55meet nicole nicholas capital one business customer and co-owner of anset uncles a plant-based restaurant and community space in brooklyn new york that got its start from a need for unity the inspiration it was born from the desire to create a space that felt like home where we can connect community culture, good food, and come together with family and friends. That's how we birthed aunts and uncles. Nicole and her husband, Mike, were fulfilling their dream of bringing people together out of their home kitchen. But they soon learned that the demand for community was greater than they knew. It became overwhelming and we were like, we need home, but not in our actual home.

19:31We realized that there was also a need in our community for something bigger in our neighborhood. So we had to find a place. Moving from a home operation into a storefront was a huge next step, but Nicole and Mike were able to take it on with the help of Capital One Business. It's not for the weak as a small business. Finding resources is super important because that's the way you'll be able to manage and scale. We would have never done that without having Capital One to be able to help us along the way. The cashback rewards are very helpful. You know, it just gave us that runway to be able to breathe a little bit.

20:05then you get to focus on the cooking of the food and making the experience great. To learn more, go to CapitalOne.com slash business cards.

20:18So we love on this show to take our listeners behind the scenes and kind of unpack a little bit, like how do you build AI products? So I want to go there next. Like how do you train these AI agents? What data do you use? Do you get access to customer data? So take us behind the scenes a bit. So what happens right now is that I think we've been, society, civilization has been changing to be more and more comfortable in recording and capturing the information. But the reality is that we make zero use of that data, right? We've been only, the most that we keep track is money and maybe page visits to your product.

21:00Or we track these datas in silos, right? Yeah, but we just track very basic quantitative stuff. And so what people haven't realized is with the advent of LLMs, you can start analyzing both quantitative and qualitative data together. And so that's the first step. I'm, you know, spilling my alpha here, right? And it's like there's all this information that is sitting there and people are making no action whatsoever. So what I found out is that the LLMs in this world of customers are really already more powerful than we need, you know, to do the work. It's the creativity of what problem we're solving, how do we deliver it, put it in the hands of the existing users, and then more importantly, how do we put it in the hands of the end customer.

21:47This is like an innovation product development problem and salesmanship and be able to go and talk to customers and assure them that it's okay for them to give you their data. Those are problems that have always been there. And people are becoming a little bit more accepting, but I do believe that humans are the biggest roadblock to our own success. They don't want to adapt it. It's a cultural challenge, the adoption of these new technologies. AI is completely different. And when I see people saying, people respond to me, it's like, it's faster for me to use Gmail and just reply to my customers.

22:25It definitely baffles me that people don't see the potential and the opportunity to do more, right? And to do it better. Well, we were talking offline about the importance of context too, right? Like when you've got all this data, how do you, that's what humans kind of do well. I mean, if you're really paying attention, you kind of can incorporate the context. So how are you thinking about that? How do you bring context into the agent's answer, I guess? I think that that's like the secret sauce. I think the application layer and solving problems is where it's at. It's not simple. it's going to be a tremendous battle with OpenAI.

23:05OpenAI has made it clear they're not sitting still. They're going to go into the enterprise. They're going to go into everything. But they cannot do everything. And they cannot serve every customer like if it's their first. And so it gives room for companies like mine to be like, I'm going to go listen to my customers. I'm going to solve their problems. Because every business is going to have to transform themselves in order to survive. And so I think that what matters is like knowing what questions to ask. The AI? To ask of the AI? Yeah, to ask of the AI. And then also there's a lot of lack of like what's the context that you give to the AI?

23:45You can't just ask anything. The background, right? Is the background, is the data, is the communications with the customers, which customers, the tools, how to go and set it and tell it who you want the AI to be? what is their role for that function, for that problem, for that question. Do you think at all about like kind of the brand or the personality of the AI agent? That's kind of, I spent like 20 years of my life kind of building emotional intelligence into machines and thinking about that. Does that factor into how you... I think that... Are your AI agents nice? They're a little like me.

24:22Okay. it's uh i would say not nice but intense you're sometimes nice and sometimes nice but they're direct yeah but they're like they want to get solved the problem right and so that's something that i think people don't realize i guess i'm also spilling a little alpha here right david hols said something about he made me journey output images the way he liked interesting i mean at the end of the day, it's kind of going to be shaped by the humans building the AI. Right. That's the context, right? And so he tweaked the style and the default settings of the images coming out of MidJourney to his taste.

25:04And when I heard him, I said, wow, that's a way to create a mode, right? People use agency today and everybody asked me, can you make sure that learns the tone of this customer success manager, this sales rep, and how they speak, and how they send that message? You know what my answer is? Do you really want to use the tone of every single person that works at your company? Do you think everybody is great? Is that really what you want? And so people don't know what they're asking. This goes back to first principles. What do I want? I want one tone that is consistent with the company. And who is the brand of the company?

25:49The founder. The founder, right? Somebody, right? Somebody has to be picked to say, we want to emulate this person. And that's what makes it unique, right? And so right now an agency, like, believe it or not, it's like, it's my tone in the emails. It's the way I write my emails. and people when when when our system starts sending emails to their customers on their behalf they're always like i like this i would send this email and it's what it's my 50 years of experience in business of how do i what i'm saying what i'm asking for and so it's it's my personality put into it right and people are like when i type that in chadjibudin i don't get the same.