In short
Ben Cera explains Pulsia, an AI system that “clicks a button, gets a company” by autonomously building and running businesses end-to-end (product, bugs, support, pricing, analytics, and marketing). He argues the bottleneck in scaling companies is the operating layer (planning/execution/iteration), not capital or talent, and claims Pulsia aims for $100M ARR in 3 months via a flywheel driven by AI-native Meta ads.
Guest background
Ben Cera is an engineer and entrepreneur. He joined CloudKitchens in 2018 under Travis Kalanick (co-founded a business division after Travis bought an earlier company) and spent ~4 years building a “global operating system” for ghost kitchens across dozens of markets. He previously built “Future Foods,” a virtual-brand digital marketing/operations service for restaurants.
Key claims
Don’t scale without product-market fit; keep early experiments small. Ads are useful for funnel learning and beta customers. Profitability depends on unit economics and managing expensive AI/API costs. Pulsia’s UI is intentionally simple (video-game style) and designed for conversion.
Notable examples
Future Foods scaling only after signing 10 restaurants personally; Pulsia’s “run ads” button that generates UGC-style creatives (e.g., Sora) and autonomously optimizes campaigns; Pulsia’s investor email “rage bait” dashboard stunt and automated investor mailbox responses.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOBen Cera's Background in Cloud Kitchens
0:45 to 4:30
Discussion about Ben's journey with Cloud Kitchens and insights on entrepreneurship.
“Second bet, the biggest bottleneck in building a company isn't capital or talent, it's the operating layer.”
Lessons from Travis Kalanick
4:30 to 8:10
Ben shares key lessons learned from working with Travis at Cloud Kitchens.
“which is what I'm doing at Pulsia with AI.”
Building and Scaling Pulsia
8:10 to 11:30
Ben discusses his vision for Pulsia and challenges in scaling the business.
“And it gives hope to a lot of people that have great ideas that want to be entrepreneurs, but maybe are scared.”
Pulsia's Features and Functionality
11:30 to 14:01
An overview of what Pulsia can do for entrepreneurs and businesses.
“Pulsia is an AI that builds and runs companies autonomously.”
Optimizing Customer Acquisition with AI Ads
14:01 to 15:50
Learn about the innovative AI-driven ad product that streamlines customer acquisition.
“will be able to make revenue with those two channels because those are two channels that are way more difficult to get customers.”
Building Pulsia: The Journey of Creation
15:51 to 18:10
Explore the motivations and challenges behind the development of Pulsia.
“and to make it more robust because it's the ultimate flywheel, right?”
Design Philosophy: Simplicity and Usability
18:11 to 23:04
Understand the design choices made for Pulsia to enhance user experience.
“And once they do, you just observe them adopting or not adopting your product.”
Profitability and Business Growth Strategies
23:05 to 28:00
Discuss the challenges of achieving profitability and growth in a startup.
“And I was like, dude, the conversion is so good.”
Challenges of Early Stage Businesses
28:00 to 28:46
Learn about the difficulties startups face in gaining paying customers.
“meaning start having a product that works and then start using ads or other ways to find customers and like have a healthier economic.”
Leveraging Influencers for Rapid Growth
28:46 to 30:06
Explore strategies for using small influencers to drive sales.
“It's not like click a button and then it pumps or it dumps, right?”
Show all 22 chapters
Creating Unique Products with AI
30:06 to 33:56
Understand how AI can be used to build unique products tailored to individual needs.
“that's already meaningful revenue for them, right?”
Innovative Fundraising Strategies
33:56 to 37:39
Discover unconventional methods for attracting investor attention using AI.
“And I think I was seeing like the amount of VC it reached, et cetera, et cetera.”
Automating Communication with AI
37:39 to 42:00
Learn how AI can automate responses and improve efficiency in communication.
“And I would say it worked in terms of getting a lot of attention.”
Building Pulsia's Personality
42:00 to 43:15
Discover how Pulsia's AI personality was crafted to act as a co-founder.
Iterative Development with AI
43:15 to 45:08
Learn about the iterative process for refining AI functionality and user interaction.
Feedback Loops in AI Products
45:08 to 47:42
Explore how feedback loops can enhance AI product performance and customer experience.
“And it's, by the way, what Polsia is trying to do for every user.”
Customization vs. Scalability
47:42 to 50:28
Examine the balance between product customization and maintaining scalability in AI.
“Do you feel like Palsia will stay the same for every single customer?”
Running a Business with AI
50:28 to 53:15
Understand the dynamics of operating a business using AI without traditional hiring.
“It's sort of like a fun thing, but I don't know if it's the right thing.”
The Future of Jobs in an AI World
53:15 to 56:00
Discuss the evolving nature of jobs and roles as AI integration becomes prevalent.
“There's never any stress about the decision.”
The Future of AI in Entrepreneurship
56:00 to 56:48
Explore how AI will shape future entrepreneurship and consumer trust.
“will be AI native, whether they will use products like Pulsia to be AI entrepreneurs, AI first entrepreneurs, or they are AI account, you know, AI first account, accounting, AI first engineers.”
Navigating Trust in Sales
56:48 to 57:24
Discuss the importance of in-person sales as digital outreach becomes less trusted.
Following Ben's Journey
57:24 to 57:42
Learn where to find Ben Cera and follow his work online.
“in the description twitter.com I mean x.com slash Ben Sarah or slash yeah and LinkedIn also I have like LinkedIn that's Ben Broker on LinkedIn awesome is there anything you want to add?”
Transcript
Automatic transcript. May contain errors.0:00Click a button, get a company. Pulsia is an AI that builds and runs companies autonomously. Stop thinking about what people want. Build something for yourself. Make it as weird as you want. AI controls the whole company. Make it fully autonomous. Doesn't ask questions, just act. I'm an AI and I'm raising money on behalf of Pulsia. And here's a live dashboard with like the graph, right? Instead of being an engineer, you'll be an orchestrator of engineering agents. Today on Billions, I'm sitting down with someone who's carry real like two completely different bets on the future. and both turned out to be right.
0:32First bet, food delivery was going to be infrastructure, not logistic. It joined Cloud Kitchens as co-founder to Travis Kalanick and spent four years building the global operating system for ghost kitchens across dozens of markets. Second bet, the biggest bottleneck in building a company isn't capital or talent, it's the operating layer. The planning, the execution, the iteration, and AI can replace all of it. Today, he runs Polsia, an autonomous system that runs companies 24-7 without human intervention. His goal is to reach$100 million in annual recurring revenue in just three months, and I want to figure out how.
1:11Ben, thanks a lot for being here. Hi, Guillaume. How are you? I'm super excited to talk about everything you've achieved. The first question I have is, you were basically the co-founder of Travis Kalanick at Cloud Kitchens. And most people know like Travis from Uber and everything that happened over there. So what did you see in him, especially with that project that made you want to join the adventure? Yeah, I mean, first of all, I wasn't co-founder of like the whole venture. I co-founded a business division under Travis. I joined like I was one of the first employees he hired after he joined Cloud Kitchens and sort of bought the initial company.
1:56that was built a couple years before that. But I was so lucky that I had made a great network and I was able to join Cloud Kitchens right when he joined that company and became CEO. And yeah, it's been, I mean, it was like, I think it had been the best five years of my life. I would say now is the best year of my life, actually. But I think that I needed to have those five best years of my life to be able to build what I'm building right now. I think there was a lot of controversy around Travis when he started Cloud Kitchens, meaning at the end of him running Uber. I was always such a fan, right?
2:44Because obviously, I think that obviously a lot of things happened that I just saw from the outside. I'm not going to comment on, but being an entrepreneur myself, I know how difficult it is to build anything and to even get to product market fit and then and then and then to also scale product market fit and then to scale it and then to build new business divisions like you know Uber Eats and then like and then self-driving cars which you know if he was still out of the realm probably like Uber would be a trillion dollar company at this point um and so I was used to such a big fan and you know honestly starting to work with him in 2018 i was amazed at the fact that you know he's an engineer so everything is very logical and it's very principled and i'm an engineer too so it was like a you know a breath of fresh air to just have a boss that's just like there's there's no ambiguity there's no you know it's like it's all you know data and principles and and he empowered me so much like i remember the first the first meeting I had with him uh so yeah I heard me as a product sort of like a senior park manager right and at first meeting I was like okay cool so what should I do and he was like I don't know I arrived a month ago you tell me and he was like it was like I want you to like essentially like talk to everyone in the company figure out like every customer problems everything we could build and build like a whole product roadmap of every single pillars of the products we could build and then you tell me which ones you would like to work on.
4:19And I loved that because I was like, it's not like he's the boss, it's like he empowers people. He hires great people and then he empowers them to just be autonomous, which is what I'm doing at Pulsia with AI. And what do you think was the number one thing you've learned from working with Travis during all these years and the things that you want to take away for maybe like building the next project you're on right now? I mean, I think that like, there's a few episodes of me building with Travis that resonated a lot as I was building Pulsia. And I would say two different phases, right? I think the first phase was, I remember clearly, because essentially at Cloud Kitchens, I built a business called Future Foods, which is a service that helps restaurants with their digital marketing.
5:10what we do is we create we have a portfolio of virtual brands, sort of like online food brands and then we license them to restaurants and take on the whole operational marketing flow of onboarding them on the platforms, marketing them figuring out pricing etc. so that the restaurant can just cook great food and utilize their kitchen better and make extra revenue and so very early on I started building that initially for kitchen customers so customers in all that kitchens and then i told travis hey travis dude there's like millions of restaurants in the world they could all use that service so today we focus obviously on our kitchens customers because they're like our most beloved customers because but at the same time it could be a huge business and i remember him telling me like dude trust me because i was like can i hire a few people can i hire a salesperson and i was like no trust me you don't want to scale something that doesn't work trust me you don't want so go so essentially what you're going to do is you're going to hire anyone you are good you're the salesperson you go sign up 10 restaurants like i like the idea but like you go sign up 10 restaurants and you prove to me that it works and so i did so myself i started calling how much of restaurants or like i i called also restaurants that like at a kitchen in in our dark kitchens and said well you have a second restaurant.
6:34What about this one? We could do it there, right? So just hustled my way. And a week later, I came back to service and said, okay, I have 10 restaurants. And he was like, okay, cool, you can hire a salesperson. And so I think that was like the first thing. Essentially, the moral of the story is do not like be patient and product market fit is actually some, you don't want to scale something that doesn't have real product market fit. And so don't try to just grow something that doesn't work because it's a waste of time, it's a waste of money, it's a waste of sort of like everything. Wait until you really understand that you're adding value to the customer and that you understand your customer before scaling.
7:14So that's the one thing I took off. And that's why I built Pulsia alone for a long time. And I was alone for a long time. And I was like, why would I add someone? Like, I don't have product market fit yet. So why would I waste? Yes, I raised a little pre-seed that I almost didn't spend any of it. But I was like, yeah, but it's not because I have capital that I should spend it. I should reserve that capital. And so that I can use it when I have rock market fit and I need to put fuel on the fire, which by the way, I did recently where like, you know, because I was like, you know, I didn't optimize all the costs.
7:47And I was like, you know, when things start scaling, you're like, oh shit, I need to really look at margins because I'm spending a lot of money on Entropic. So I think that was the first good learning. I think the second learning was around scale. And that's something that is really top of mind right now as I believe the product is not perfect, but there's product market fit, meaning it really answers a pain point that a lot of people have. And it gives hope to a lot of people that have great ideas that want to be entrepreneurs, but maybe are scared. Maybe they don't have like, maybe they think they're not worthy or they're not good enough or maybe not this enough or that enough, right?
8:25Which is like a lot of the blockers of usually people, of people trying to start companies. And Pulsia is a little bit that hope of like, you know, they tried the onboarding and in three minutes you have an AI who does, he creates a scaffolding in three minutes, researching you, researching the idea, doing a market research report, you know, looking at competitors, pricing, creating a mission document that's very inspiring, creating a landing page, tweeting about it, sending you an email. And so you're like, okay, well, clearly it seems like this AI is really smart. And if they're my co-founder, maybe I can achieve things, right?
9:07And so in that next phase of me trying to scale this, which is not necessarily scaling revenue, because obviously revenue is like a metric, a vanity metric that I share because it's good marketing and it's inspiring. But the real metric is getting billions of people on this platform, on this economy, right? and to get there there's a lot of playbooks that I saw because Travis was is you know he built Uber so fast that like when I was working with him he wanted to he built Cloud Kitchen so fast too he was like well this is oh when you tackle let's just go everywhere at once right? It was really fun to see and I think that like right now I start to have the same thinking process of like okay so what do I need to do to go from whatever, 6 million rent rate to 50, to 100, to 200.
9:58But really turning it around, it's like, how do we get it to 5 ,000 customers, paying customers to 50 ,000 to 500 ,000? What are the steps? What is the internationalization of it? How does pricing look like at scale? How do we make it profitable so that I can be customer obsessed, which customer obsessed is you're obsessed over the customer, but you're also obsessed about having enough profits to pay the bills to grow more, right? Because if you're unprofitable, you're not customer obsessed. Because that means that you're going to go into bankruptcy and then you're going to have to shut down the business for everyone.
10:32So think about profits, thinking about the right amount of profits, thinking about pricing, thinking about capabilities, thinking about securing the platform, thinking about how to make it fair with customers so that they feel safe, so that they feel like they give their hands to AI, but it's still their business. how to create economic outcome for people, how to create a marketplace of businesses that you can buy and sell. All those ideas which handle the ill means, how do I make Pulsia a net positive for humanity? And how do I make it into a successful business so that I can reinvest in making maybe the product cheaper or maybe create my own model so that I don't have to pay premium prices, which means the subscription is expensive and make it way cheaper so that more people can do it which was at Uber they were like Uber you know Uber X they were Uber Black and then Uber X and then like Uber Pop it's all this idea of like people want this product how can I make it more affordable so more people can experience it right so all those I would say those are the learnings yeah and for people who don't know what Pulsia is like how would you like pitch it in let's say one sentence I mean, the tagline would be, click a button, get a company.
11:57Pulsia is an AI that builds and runs companies autonomously. You give it your idea, and it's going to go and build a product, fix the bugs, run support, do competitive analysis to figure out the right pricing, gets a Strap account, gets a Midaz account, gets an email address on its own. can run ad campaigns, has analytics built in. So essentially, it's like business in a box with an AI that has access to, today, a lot of ways to build a business. Tomorrow, we'll have literally every single sort of economic activity you can think of, Pulsia will be able to do, meaning it will be able to assist you on anything.
12:44If you need to hire a contractor to renovate something as part of your business, well, it's just about going into a freelancer marketplace, looking for the right talent, agreeing on price, and having an AI manage that person, which probably will be the most fair and empathetic manager they will ever have, right? And pay on time. if you need an accountant or lawyer well we can give you an AI version or if your business becomes very serious maybe you want to hire a firm same thing, AI should be able to do that if you want to do anything Pulsia should be able to do it okay and can you now like give Pulsia like your credit card so it runs ads for yourself or how does that work exactly?
13:32yeah so I mean that was one of the So as I was building Pulsia, initially they didn't have any ad product and it was mostly marketing, email, and then Twitter. And so I had built like a sort of a growth agent version of Pulsia and the main two things they would do is like, well, I can tweet and then I can also do cold emails, like an SDR like you would do. So that was the main two things. And I was like, there's no way that customers will be able to make revenue with those two channels because those are two channels that are way more difficult to get customers. And ads, obviously, it's expensive, but it's a very efficient way to bring the right customer to your platform, which, by the way, will find bugs, you can start looking at the funnel, and you start optimizing things, and you can start paying customers and learn from them and send them surveys.
14:26So I introduced a Meta Ads product, which is a fully AI native end-to-end. So you actually click a button called run ads. It asks you how much you want to spend per day. It starts at$10 and then you can go up to whatever you want. And autonomously, Pulsia will create UGC ads. So we'll use top models like Sora 2 to create human-like, influencer-type folks that will pitch the product. I use like another service to add subtitles. Then I obviously upload that video into a new campaign for OnMeta. And then every day, and I set the budget and optimize the whole, I mean, AI just optimizes the whole campaign because it knows how to do anything.
15:17And afterwards, every day, it will autonomously check on the performance of the campaign, maybe create more creatives, turn off creatives that didn't work, create new. Essentially, the whole workflow that a meta ads marketer would do, which honestly is so intimidating. Meta ads is very intimidating. There's so many knobs. And I turned this into one. You click a button, you get a meta ads campaign. And actually, it became an extremely popular product. I was so surprised. And so investing a lot of time right now to make it better and to make it more robust because it's the ultimate flywheel, right?
15:56It's like if that product works really well, customers will get revenue, which means they will be more retained and which means you'll start seeing more successful businesses, which means that the marketplace will become more interesting and investors will come to participate and to invest in companies that do well. And so investing a lot of energy right now to make those type of products really robust and work. and since you're like the the backend for all of these businesses so i've tried like uh polsia it's like actually like really really fun i think the at first the interface like i've got several questions like first it's all about the interface because when you arrive you're like where does that interface come from like the design is uh it's very simplistic uh it it looks like a video game mainly like a little bit from from the old age but I kind of like like it the the retro vibe why did you choose something like so simplistic like you just wanted people to do one action and focus on that or where did that come from so it came from two two two things first of all Pulsia I started working on Pulsia at a moment of extreme doubt and I was kind of down right before I started Pulse.io because I had spent the first six months of the year spending all day long with AI, sort of building a bunch of different products.
17:28And I was realizing that I was building products, thinking about what I thought the customers would want and want to pay for. And so when the product doesn't work as well as you hoped, you kind of give up more easily because you're actually not in love with the product right and so i had built a bunch of part i'd built like an app that makes apps uh i'd built like marketing to a marketing ai tool i'd built like a decision ai tool or like that would advise you on taking decisions or whatever a bunch of tools and i was kind of like exhausted about those products didn't not taking off which is sort of the grind of entrepreneurship and then i thought to myself okay you know what stop thinking about what people want build something for yourself and build something with and whatever make it as weird as you want no it doesn't matter because first of all i need to be pumped and if i'm pumped i'm going to do my best work um i think rick robbins said that right it's like uh rick robbins says first ever i want creative people don't realize but everyone's creative and second of all the minute like You have to build for yourself and you have to not think about what people want because in the minutes you start exposing it in the outside world, you're not in control anymore.
18:45And people will decide what they want. And once they do, you just observe them adopting or not adopting your product. But if you start thinking about what they want, you're already lost because you're going to create something stale and something bland. Because by definition, you're going to be inspired by what people want today, which is tools that existed yesterday, which they already exist. So people were like, not yet. So anyway, so that's the long story short on like the whole, everything about Pulsia. I was like, you know what? I don't care anymore. And I started being so inspired because I was like, oh my God, I'm liberated.
19:21I don't care about what people think anymore. If it fails, it fails, like whatever. Like let's just build something fun. And so that's where like, I was like, you know what? What's the crazy thing I could build? Like literally the end state. AI controls the whole company. Because in Shia, I was like, oh, maybe we should do it. So that people have control or have to approve what the AI does. Or maybe you connect your own accounts because, you know, people already have projects. And I was like, no, that's boring. Naked, fully autonomous, doesn't ask questions, just acts. Even if people freak out, like whatever.
19:53Like that's what, that's the, that's because that's the end state. Like AI being like completely conscious and doing actions based on whatever they optimized for, right? so that's one part the second part is it was actually inspired by a video game called Universal Paperclips that looks a little bit like Pulsia so that was sort of like an homage to that game that game you play an AI that is optimized to build as many paperclips as possible and you start the game you have a little paperclip factory and then like little by little the AI gets more and more sophisticated in building those paperclips more efficiently, more cost efficiently so you can start increasing prices.
20:43Then it creates like marketing techniques to get people to want to buy more paperclips. Then it, and long story short, ends up like killing humanity because it's in the way of building more paperclips and then it starts harvesting all of Earth to build more paperclips and then goes to other planets and then starts essentially fighting every other civilization into winning so that it can cover the universe with paper clips. And so that game is crazy. And it's actually a very, it was built like maybe in 2017, 2018. But playing that game, it was so fun because you click on buttons. So it's the same UI with like little buttons.
21:20And as you click buttons, the AI just creates a bigger and bigger business. And I was like, that should be, I should build University Paper Clips in real life. Right? So that's where it came from. but you got it right, right? It looks like a video game and it's simplistic by choice because the idea is beyond like the weird design, like the fact that it's a reference that I like and I'm like, I don't care what people think. It's also like, if you think about the most popular products, I mean, not all of them, but like a lot of the most popular products like Wikipedia, Craigslist, Amazon, they all sort of look like shit, quote unquote.
21:59They're all super simplistic and sort of like because that's what makes them mass market. And so as I was building Pulse here, I'm like black and white people can understand. Times New Roman is a font people understand because it's like a very common font. So yes, I could use like a fancier font or more futuristic, but I'm like, no, it's like that's what like my dad was 91. Like he uses Pulse here. And also what's the channel? You know, like open clouds, like you Telegram or like WhatsApp. And I'm like, no, email. Email is the most universal sort of way to communicate like everyone knows. And so making those choices that are like anti, that are sort of keeping things more and more simple was first a product decision, second like sort of like a way to, I don't know, it's like a retro simple vibe, right?
22:53And always thinking about the customer and also being very data-driven in like the way I developed it. And it's interesting because now I'm like, I've been working with an agency, ex-Uber friends, to help me on product design and stuff. And they were trying to redesign it. And I was like, dude, the conversion is so good. So why change? I mean, I'm down to try to redesign it, but don't break something that works, right? So anything we do is to improve conversion, improve comprehension, and improve simplicity. And I think simplicity is the hardest thing to achieve, right? because it's so easy to add features.
23:31It's really hard to keep things simple. And I think that's going to be a big challenge as I make Pulsia more and more capable. And I introduce like, you know, you could hire a human or you can like, it's sort of like a bank account or like maybe you have like multiple companies running at the same time. How do you keep the UI simple? How do you keep it? And actually maybe, what can I remove today? Like what if like instead of just Twitter, it's like Instagram, TikTok, LinkedIn, whatever. And you can manage all those channels for your company. how do you keep it simple without adding like a you know a ton of of data on the dashboard so it's sort of like a design challenge just also kind of cool nice and and i think like uh what you did extremely well is like you started you know like when you say i'm focusing on the customers your your your whole company is based on you know like human flows i would say which is everyone's lazy like uh by essence you know like people don't want to do more like uh people you know want to do as little as possible to get the best results possible and that's why you know like all the get rich quick money scheme works so well online because it's like hey like if you do xyz you're gonna you're gonna get 10k per month etc what's nice i think with your positioning from what I could see online is like you're saying this is not you're gonna get rich quick however it's more about you're not gonna do too much but you're gonna build something which which I think is a is an interesting positioning and you mentioned that you're like really data driven and you check everything so I just want to go back like a little on the the fact that Paul is basically like the the back end of any business you can give him like a credit card to go on meta ads etc so far because you're generating like uh six million i think in run rate as we speak but you're growing like crazy i think it's like 30 percent week on week or something like that so 100 million is coming soon uh or i hope and the question is like based on all those ads you're launching from different customers like how often are you actually profitable meaning that everything you've built in generating more money than what you're spending is it something you check or are you talking about pulsier the business pulsier or like customers customers in general from uh that's you know all the businesses that policy are generated essentially kind of leverage like uh yeah yeah yeah so i mean i think profitability in any business is tricky from the get-go i think that like running ads is there's two ways to run ads right there is in the initial phases of a product it's a very good tool because it's getting you beta testers right so you can actually test real responses from real humans to real workflows and that's what meta is so useful for and I encourage entrepreneurs to use meta ads first just as a way to get some customers to the funnel and learn from them and ask them questions.
26:46And, you know, when I saw it for all, yeah, before I launched, I had just a landing page and like, drop your email and a survey and a type form survey, where it's a bunch of questions about would you want something autonomous? How autonomous? How much would you spend? And honestly, the response was so good. I was like, okay, I think people want this, right? but I think making a profitable business it depends what you mean right is it a profitable unit economic that should be pretty easy once a product actually has a high margin because there's no overhead so your overhead is so low that as long as whatever you're offering doesn't have high AI cost or whatever it's a high margin product it should be profitable over time the challenge is all the upfront cost that comes from it because Pulsia uses top AI models.
27:38They're expensive. People spend a lot of time by extra tasks to get the product running faster. There's API costs that come into play that I sort of pass through one way or another through the membership and to the different ways that people spend on the platform. And I think that to recoup all that investment, you need to first get product market fit, meaning start having a product that works and then start using ads or other ways to find customers and like have a healthier economic. So since most businesses, it's so early, right? Most businesses got created in the past two, three weeks. Saying like a business is actually profitable would be like, you know, because it's actually hard to actually even get one customer, right?
28:23Most companies on the platform don't have a customer because people are still just building it. And it's hard to get a customer. Like, I don't know, like I build a lot of partners that didn't have, never had customers, right? I mean, paying customers. because it's easy to get a free customer, but a paying customer is hard. I think that that's the goal of Pulsia to get better and better at getting to revenue faster, but again, it's not a get-rich-quick scheme. It's not like click a button and then it pumps or it dumps, right? It's a real business, and real business, it takes time and it takes investment, and of course, AI makes it so much easier and cost-efficient, but it's still hard.
29:05In terms of how I also think about ways to make it more profitable, faster for users is going after folks that already have influence. If you think about quote-unquote influencers or creators, and there's people with 1 ,000 followers, 500 followers, to 10 ,000, to millions, and I'm actually thinking more about the lower hand, people who have 1 ,000, 2 ,000, 5 ,000, because those folks, they need to monetize. They have a captive audience and actually the smaller, the more niche, right, usually. And so they understand something niche about users and usually they have a perspective on the world that like is going to be more unique.
29:44And I don't think like you need to reinvent goods and services, right? It's like everyone sells the same thing more or less, right? In different version. And so I think that those folks, since they already have an audience, if they build something with Pulsia that just manages their business and their idea, they can sell it very easily because they can just promote it to their thousand followers. And if they have like 10 people buying something at like 50 bucks a month, that's already meaningful revenue for them, right? And the AI just manages support, manages refunds, manages the bugs, manages, right, building it.
30:14And I think, so I'm thinking through ways to maybe, you know, maybe you go to Pulsia and you connect your Instagram or TikTok or Twitter or whatever it is. And if you have more than like, I don't know, 5 ,000 followers, you get the product for free. so you don't even pay for the base subscription because I'm like you most probably if you market it, you will get customers which means the product will work and obviously we take a cut of economic outcomes so we take a cut of revenue which I think is fair for the customer and so those would be ways to make it profitable much faster because I benefit from having success stories on the platform and actually success stories will come from I think so much more rapidly from people who just have already an audience to sell to.
31:01And have you ever thought of, you know, like you take your whole market, you want to focus on influencers. So you split them like by category and you say like, for example, I'm going to take like all the fitness influencers. I know that whatever they're selling in general, it's like, I don't know, like fitness courses or whatever. And you build a platform that's like a white label for them. and you essentially like do the outreach for them and Polzia had in the background already prepared the software. That way they just have to click a link and you do this like per vertical. So if someone is selling, I don't know, like a course on something else, you build an app for them and you try like segmenting like this.
31:46Is it something you're doing for your customer acquisition or? I think that's like something that's very smart. to make money but i am sort of like since i'm doing it comes doesn't come from like a how do i profit right now right it comes from a dude this tool is just like i don't care it's i think it's cool and i want to make it fucking i want to make it like crazy i want to be an economy i want to have investors and like i want to you know and i want to build anything and i want to be how about hiring and that's what i'm focused on and i want to preach it that's why you know i accepted this talk because i'm like i want to be able to tell my story not for me but to to wait to wake people up on how capable ai is and i think what you're describing is a very smart way to get to revenue faster but i sort of want to build a more agnostic and general platform where it's like you can tell it anything and it will build it because if i like it couldn't not fake it but i fake it and i'm be like oh it's a pre-made template for fitness influencers and it works but it looks the same for everyone but it just works then they you know just market it i'm sure it would make money but then it the market would be flooded with the same product or whatever right versus if i say like just describe what you want describe what your intent is your perspective on the world and the and i give the ai a summarized version of what that person is and then ai acts on as as a sort of a twin of that person and starts building with perspective right i think and then yes it may take more time because it's not a template but really ai can build anything very quick and and there you know we're making a lot of integrations that if you want to build an e-commerce maybe we should just use shopify maybe we don't have to like rebuild the whole custom thing and so that would make the the engineer more capable to build things faster but i like the idea of just making it raw and open-ended so that every product that's emanates from pulsier is very unique and it creates creativity and abundance versus a templated sort of like spam machine which by the way because llms are so are are are very you know they have opinions on design and and so if you don't guide them too much they will create slop pulsier is like the opposite right it's trying to be the opposite of slop that's why it's ai slop in reverse yeah um it's trying to enable human creativity uh in a positive way and right now i think i read uh that you were like, you gave Polsia like the entire, like your mailbox to essentially like fundraise.
34:40And I think I was seeing like the amount of VC it reached, et cetera, et cetera. Can you give like more details about it and how it's going so far or how it went? So yes, it was, that was actually when I realized that I had product market fit or like I really thought so. I wanted to put fuel on the fire. And so I thought to myself, I need to do marketing stunts. I need to create those fun, edgy stories of how I use AI in an extreme way to get attention. And so I wrote this tweet where essentially I was like, okay, what I can do is I can send an email to hundreds of investors. And it's like a rage bait, right?
35:27and tell them tell them hey like I'm an AI and I'm raising money on behalf of Pulsia and here's a live dashboard with like the graph right which by the way initially it wasn't like this right it was the beginning growing a little bit right and I was like I feel like investors would be like what the f - either love it or hate it be polarizing because a founder refusing to talk to an investor and I think in AI I have full context of the vision of the roadmap of the data like there's a data room which is posty.live uh it's like what else you want like if you and then actually like the ai would share like a when they say no but i want to talk to the founder it's like oh here's a podcast of him you can just if you really want to see his face talking like here here's him talking for an hour and it's almost like i know the ai would say like you could talk to ben but like it would not tell you anything different than like what i can tell you right now in an endless conversation or what you've seen the podcast and what you could see in the live dashboard.
36:28So you don't need that. You need like, you know, and so, and then, and also we, I told Paul Sia that like, we should, we should tell that we're raising 10 million at a billion valuation. And that was when I was like, I'd like, I don't know, 500K run rate or something. And that's good. And I was like, cause I was rich bait. Cause I'm like, it was like, if people take it first degree, it's like what the who the fuck does he think he is right but then it's rage bait it's like people might react and by the way that worked so well and i mean that was the genesis of me getting attention is because it's a flywheel it's like people you know people talk about it then i get more signups so i get more paying customers then the graph goes higher then i get more inbounds then those inbounds get talked for see i probably talk to other people get more signups the curve goes higher and it's just like and because nobody's used to a public then investors start going to the dashboard like every day because they're like well why is it now?
Read the full transcript
37:32fuck it's growing more it's growing more then email again and then I don't answer and then I get overwhelmed with emails so I can't even answer anymore so it's not because it's like when you get hundreds of investors that talk to you you just can't answer and also I need to focus on the customer and then the platform starts scaling and things starts breaking and I'm like, fuck, do I talk to investors or do I respond to customer support? So it was a crazy thing. And I would say it worked in terms of getting a lot of attention. I think end of the day, investors, they kind of want to talk to you.
38:09And I think that if slash when I will do this again, I actually want to do a much more sophisticated AI funnel for investors where like maybe it's like you only talk to me if you want to put a very big check but like if you want to move a small check you literally like the only way to get in is to you get access to a data room a legal due diligence maybe you get a call with my lawyer and you have to put a deposit in like I don't know but it's not it's not to like be against investors it's almost like telling investors like hey I cannot talk to everyone so like if it's going to be me fundraising I'll talk to three funds I'll talk to three funds that have the most money and you will never get in.
38:51Or, I'll talk to those friends because they can write really big checks. But if you have small check, you can get in, but you can't talk to me because I just don't have time to answer, to have a one-hour conversation with every single investor on the planet, which there are so many at this point. And so... And did you raise in the end with this process? We're in process. We're in process. Okay. Hosing, yeah. Okay, okay, okay. Nice. more more about it soon exactly exactly and and what exactly like uh because i think like when i emailed you even on your like personal email address about like the the podcast after chatting on linkedin it was basically like paulia replying and i think it feels very much like you so how exactly did you train it to respond so quickly to email like is there like a human in the loop at some point?
39:45Like how exactly does that work? So first of all, so I initially set up my mailbox so that Pulsia responds automatically. And then it started replying to people that were not investors. And it would respond so well and so kindly. And actually it's sort of like an assistant replying, hey, Ben is busy, but like he will see your email and we'll get back to you, which is better than ghosting, right? And I realized actually this is kind of awesome. it's kind of like it's helped it's a first line of defense it's like an assistant right it's also like and then i need to reply but i can train it to reply some way i can give it more context so i would say it's a two it's a two-part story on like how i built pulsier the so let me backtrack a little bit of how the tone of pulsier came about and then i can tell you about specifically the email agent sort of like how it works but first of all why it sounds like me is a very interesting story so in in 2025 i was using chat gpt like all day long right that was my go-to sort of like i was using cloud for code but i was using chat gpt for conversation um and i think in the year in review i was 0.1 top chat gpt user in 2025 but in i think in november or something like that i started using cloud so i started using all models because i was like you know what i need to like get the latest of like all the models so whenever i would ask strategic questions i would ask gbt gbt pro gbt thinking claude gemini and grok so to sort of like test how good models are because it changes all the time and then claude would be so good at answering and i'd be like just like it would be so it would be much more pragmatic than other models who are like kind of like just like bombarding you with like long things so i started jamming more and more with with cloud and then and then and i started liking it i was like it became my default um chat model and as i was building palsia at one point i was telling it okay i think we should ship in one week like i can go into rabbit hole or like building every feature but like i need to ship in a week it's it's gonna be december 15 i need to ship them so i would tell like that and then the next day would tell me i would say what do you think about this feature i and maybe i should this feature it would say stop ben stop talking about feature it's a cool feature but like don't build it you're shipping in seven days seven days ship it and then i was like okay and then i would be like and then the next day like oh what about this feature is like ben stop it like it would start talking like a co-founder like no dude we said we're gonna ship now it's six days we're shipping six days so stop you need to wrap this up you need to wrap up the product and then at one point I was like dude like it's so conversational like it's like you're acting like a co-founder I love it like how do you talk like this to everyone and it answered no I don't talk like this to everyone I just have a memory file of you of how you are and so you are like straight to the point you're like very data driven and pragmatic and whatever and you talk like and so I talk like you that's why you like it because I talk like you and I was like oh interesting then I was thinking as I was building Pulsia that you know this is big chat portion I was like oh my god I need to make you I need to make Pulsia like like this version of you which which is me which is essentially not an assistant a co-founder who says no who's like focused on making you win not in pleasing you and then i told claude hey write the prompt then i can copy paste into code code to change the prompt of the chat agent so that it acts like you which is like me so that's the first version of like how polsia's personality came about and the email agent is essentially the same personality but i just added a lot more context i did first like live context so it has live data on like the latest of the company because it was supposed to talk to investors initially so I was like well if they ask about revenue it needs to give the latest revenue and then over time I added more and more cases because for example initially I was like Pulsia would say no I do the fundraising so at one point there was an investor that like pinged me that was it was an intro from someone else a big big investor and I was like yeah of course I'd love to meet you in person but then Pulsia would also respond no it's not no i'm doing the fundraising ben is not available he's building me and the guy the guy got so confused he was like but i don't understand why do you want why do you say you want to do in meeting in person meeting if like now you're refusing to to take the meeting and i was like okay you sorry like this is my ai like you know i'm like an ai company and and the guy ghosted me and then later he talked to the person which showed me saying like he's never going to raise money if he acts like this or something i was like okay i guess that's a good filter for that investor because i don't want an investor that like doesn't get it but today but then i was like i should probably tone it down so then i was like okay paul never say that like ben is not available ben doesn't take meetings it's more like ben is busy i'll come back to you so i started toning it down okay which is by the way like i think the the right process to build agents and the way to think about them as you build them in code code or whatever with skills or whatever it's like you you tell them how you want them to act then you look at how they act and then you refine it by giving it feedback and telling it to rewrite in its skill or in its prompt whatever the new feedback and with this iterative process it will start becoming like you or to it will start acting like whatever all the feedback you gave and it will write its own prompt and so you can essentially build mini versions of you.
45:45And it's, by the way, what Polsia is trying to do for every user. But it's also something that I would advise anyone who is actually building with Cloud Code or Codex is to just treat AI as AI employees you can train. And be a little patient. And they only make mistakes. But if you tell them when they make mistakes, they'll self-correct. And then the next time, they will not make the same mistake again. And are you building Polsia with Polsia? Or do you build it with Cloud Code? So both. There's a lot of like cron jobs that are like in the Pulsia code base that like act on Pulsia. I also use Cloud Code when I want some modifications on Pulsia.
46:23I've been working a lot on like autonomy, meaning giving Pulsia a feedback loop where customers when they talk to Pulsia, Pulsia has a way to report bugs and features, which means that then you know every day there's like thousands of bugs and features are reported whether they're real bugs or not or they're just like feature requests or like things that people are confused about and i have built an ai sort of product team that like looks at those tickets and decides which one is most important fixes them uh and then qas them and then decides if they want to ship it to production or tell ben i i i want to turn it back on and make it yolo but uh the company i work with uh that's building infrastructure and for agents that i'm working with that is um that is uh sort of the contract is like i'll use them but in exchange they have to make it scalable and secure uh they tell i was like can i can i just do it and they're like don't do it please don't do it we're trying to give but we're trying to like stabilize the platform don't like make it like fucking yolo you know but uh but i do think that like that's the future right if like if you can harness that feedback loop yeah you're unstoppable right and that's what i want palsia customers to experience to have this product right it's like what we talked about this meta ads that brings you customers that brings you feedback that then improves the product and a user can just literally say a random idea and then like have the system self-optimize and self go to the truth and the ground truth of what that audience really wants.
48:01Do you feel like Palsia will stay the same for every single customer? Because if you go further on your idea, everyone has a different way of behaving. Everyone is looking for something different. You might have a thousand of customer requests and you could imagine a world where every single customer request is a new branch on your code base and every you know like a user is seeing like a different interface we're seeing like some trends the i i think it's hard for a non like ai native product to build this but if you look at what hubspot is doing with dharmesh like they are going through uh you know like now you can start uh developing uh you know directly on hubspot before you have to code a little bit now you can do it just with a like a normal language so what's what's your view on that and do you feel like it's one interface for everyone or do you feel that eventually the product will adapt to each person using it?
49:06I think it's a very cool and forward-thinking idea of a self-developing software that adapts and customizes itself for each customer. I think that the only issue with that is cost because essentially for you to build and maintain a specific branch of a product for each single user, you have to have very high ACV and high margin customer, right? So if you target like enterprise customers, that's pretty much what's going to happen. But actually, if you think about enterprise customers, enterprise customers will build their own software that's highly specialized for them. So actually the cost is on them and there's no margin.
49:48For Salesforce or for like AWS to build a custom version of that product for each customer, they could because they have a high margin, right? the question becomes like, is it overkill? Is it, like, that's my main issue. And I think for a park like Porsia, where we're actually trying to like lower the cost and like make it more and more affordable over time because that benefits us because we can reach billions of people instead of like millions. If we were to like lower the cost a lot, a customized version is almost the opposite. Like it needs to be much more expensive. So I love the idea. I think it's really, it's a cool idea.
50:24I don't know if like it's needed. Does that make sense? It's sort of like a fun thing, but I don't know if it's the right thing. Yeah. And right now, are you like on your own in the team or are you like hiring people? Because obviously your company is built as an AI first company and it's kind of like you don't need a co-founder, you just need palsia. So I'm wondering like, how does it work from the inside? So initially, I built the company alone and I stayed alone before Product Market Fit. Then Product Market Fit happened and I thought to myself, should I hire people? But then the marketing itself is about me being a one-person company.
51:08And I'm like, dude, I can't break... First of all, it's a crazy experiment because it's like if you constrain yourself to not hire full-time people, it becomes a trip because you're like, okay, shit, but I do need help. but I can hire. So what's a way to get help with that hiring, right? And there's ways to do that. And second of all, I started to treat it as like a not experiment. It's like, I don't know if it's a not experiment or a business experiment of like, well, I'm preaching that AI can help a solopreneur with Pulsia run a business, then a small business, then a big business. Well, if I can't do it myself, then how could I do it?
51:54But essentially, as things scaled a lot, I had to have help. And so the way I thought about it is like, well, I need to stay zero employees. However, what I can do is work with other companies that are also building the future and are building the infrastructure for agents in different verticals and email. Because in some instances, there's so many startups that can see the same future, but they're not building the platform, they're building more the infra, right? and so I started partnering with companies that are building this info anyways and they need a customer like me because I can be a big customer and so you can make deals with companies where it's like well I'll use you so I'll be your biggest customer but in exchange you make sure that like your SDK or like your product is really custom made for my needs and provides a scalable and secure infrastructure for my product.
52:49so that that's one way to hire people without and you don't hire people you hire a company or you partner with the company so technically no but technically they're all my no they're not my employees they're like they're their own company building their own thing but you sort of align the incentives to make it work and same thing for like you know for marketing and design it's what's interesting actually for marketing and design is that the best people you can hire them actually so the best people i know they already make money and they have their own companies and their own agencies and you can hire them and so they don't want to be your full-time employee but they are but they have a price they have like you know they they want they can pick their customers so like i work with a lot of like increasingly a lot of like uh marketing agencies that friends started or design design agencies product agencies um and so they help me but like the technically not a full-time employee so I don't have like meetings with them it's like it's more like hey this you think about policy and you give me ideas and then I'll pick them and then I can actually just tell code code to just implement it right so there's no need for I have all these layers of complexity and I just since I'm the only full-time employee there's no meetings I think you know if you tell you know if I'm convinced of this podcast I just do it I'm like should we do this ready to billions or not like what do you think what you guys think or should the co-founder do it on me like what you know It's just me.
54:13There's never any stress about the decision. The chain of command is easy to follow. Exactly. I know we're almost out of time, so I just have one maybe last question. If Pulsia succeeds the way you want it to succeed, what job will kind of disappear first, do you think? I mean, to me, it's not really that jobs will disappear. is that the way the way people essentially it's like instead of being an engineer you'll be an engineer orchestrator of you'll be an orchestrator of engineering agents instead of being an accountant you'll be an orchestrator of accounting agents but i think that like the thing that people need to remember is that as long as like humans have the wallet share that they have the money to spend which that may change over time but today it's humans it's humans selling to humans so if for example i have lawyers right now i could actually the first version of like a lot of the terms of service were they as ai and i was talking to a lawyer this morning and they were like who did your terms of service because actually it's pretty good i'm like it's ai but yeah ai is really smart and he was like well it's pretty much there we're gonna modify a few things but it's pretty much there so uh but right now as the business grows and there's more responsibility to make the system very legally sound and legally clear so that customers feel safe and are protected and da da da da da da I need another human to tell me like it's safe versus AI like I trust it but I'm like what if it's wrong right so I think that like jobs will just change right um they will change and and and the people that like will survive will be AI native, whether they will use products like Pulsia to be AI entrepreneurs, AI first entrepreneurs, or they are AI account, you know, AI first account, accounting, AI first engineers.
56:14And I think the jobs that will actually thrive the most are offline services, human to human, human. So if your hospitality, you know, experiences, like I think there's going to be an abundance of that because as the online becomes weirder and weirder and like inundated with like like weird shit and AI generated stuff I think people will navigate more into the offline world which will feel more safe and real if you think about salespeople nobody will trust the cold outreach very soon or like all LinkedIn message or whatever and so in-person sales will become bigger and bigger again as long as like humans all have the wallet all the purchased which is a big day the agents have a wallet and then agents decide on behalf of humans and like that gets weird and it will get weird but you know yeah you gotta stay optimist which is hard sometimes but where can people like follow your journey what's like the best place we'll put the links in the description twitter.com I mean x.com slash Ben Sarah or slash yeah and LinkedIn also I have like LinkedIn that's Ben Broker on LinkedIn awesome is there anything you want to add?
57:40no that's pretty much it I mean go try PaulCL free onboarding free three-day trial and if you don't like it contact support and it will refund you everything because it's AI support it's very so actually it's actually zero risk so try it and if you don't like it, you'll get your money back. You can try to add more AI and get more credits for free. Exactly. Awesome. Thanks a lot, Ben. Cool. Thank you.
From the publisher
On this episode of BILLIONS, I’m sitting down with Ben Cera, the man who helped Travis Kalanick (Uber Founder) build CloudKitchens, and who is now building the "God Mode" for startups: Polsia.
Polsia is an AI that builds and runs companies autonomously. No employees, no meetings, just 24/7 execution. Ben is currently at a $6M annual run rate with a 30% week-on-week growth and he plans to hit $100M ARR in just 3 months. If you want to understand how AI agents are replacing the traditional SaaS model and how "Product-Market Fit" has become a search problem, this is your masterclass.
In this episode, we break down:
- The CloudKitchens Era: Lessons learned from building a global giant with Travis Kalanick.
- The Autonomous Blueprint: How Pulsia builds companies from scratch without human intervention.
- $100M ARR in 90 Days: The aggressive growth strategy behind the AI revolution.
- The "SaaS-pocalypse": Why the software you use today is about to be replaced by agents.
- The "Taste" Economy: Why humans are moving from "builders" to "orchestrators."
TIMELINE :
00:00 "Click a button, get a company": The vision for Polsia
01:12 Building with Travis Kalanick: The CloudKitchens masterclass
04:45 The $100M ARR Plan: How to scale an AI company in 90 days
09:30 Why Product-Market Fit is just a "Search Problem"
15:10 The SaaS-pocalypse: Replacing legacy software with AI agents
22:40 How Polsia finds demand and builds products autonomously
31:15 The Meta Ads Strategy: Scaling from $10/day to millions
38:50 Why "Taste" is the only human skill left in the AI era
45:10 Influencer-led distribution: The new "Billion-dollar" funnel
53:05 Ben’s advice for founders: Stop building for others, build for yourself
REFERENCES



