Sharath Narayana: 06:33
The first time, right? Yeah. Yeah. And now it’s just a mobile app. But what my team is thinking is, hey, mobile app is cool, but you still have to open the app. But hey, what if it’s just a Chrome extension? Where or extensions where people are like you and me are talking, or people are just listening to your to your podcast of me and you speaking and but they want to hear this in Mandarin. They want to hear it in Japanese. They want to hear it in Spanish. And, and how do you do it in real time where they still feel it’s Sharath and Dr. Jeremy speaking.
You still keep the identity of both our voices, but it’s just that I’m speaking in Mandarin and you’re speaking in Mandarin, right? So that changes the whole complexion, right? And it was the first product we built with a very mobile first interface. Up here, it’s just a demo that you’re seeing on your website where you can just see the demo, but you can download it. It’s on App Store, both on Android and placed in an ice stores. You should download the app and use it. It’s free. I think it can be a game changer.
Dr. Jeremy Weisz: 07:33
And we’ll get into it. I’m going to formally introduce you in a second. This episode, by the way, is brought to you by Rise25, at Rise25, we help businesses connect to your dream relationships and partnerships. We do that in a few ways. One, we’re an easy button for a company to launch and run a podcast. We do the strategy, accountability, and the full production execution behind the scenes. Number two is an easy button for a company’s gifting. So, you know, staying top of mind could be clients, partners, prospects, even staff from a culture perspective. And I tell people think something to send them your partners every four months for five years, right? And we don’t like sending tchotchkes. It’s like actual food because people enjoy getting food. Most people do. At least I do. So we kind of call ourselves magic elves that run in the background to help people build amazing relationships. And that’s the number one thing in my life. I’m always looking at ways to give to my relationships.
I found no better way one to profile the companies I admire in the podcast and share with the world what they’re working on to send them sweet treats in the mail. So go to Rise25.com or email [email protected].
I’m excited to introduce formally Sharath Narayana and he’s CEO, Co-founder of Sanas, which we saw. He brings over 20 years of enterprise software sales experience and AI expertise. And he’s been a serial entrepreneur investor. He co-founded Observe.ai and invest through Carya venture partners. And he’s really like, as we said, if it’s speech, it’s honest and it’s a real time speech layer for the modern world. So it really makes conversations between people and conversations with systems clear, natural and scalable, which we’ll go into. So thanks for joining me and just talk about Sanas and what you do, and I’ll pull it up again.
Sharath Narayana: 09:20
Yeah. And thank you so much for hosting me. It’s always amazing to meet, meet fellow entrepreneurs like me and, and talk about each other’s journey. And I always come back learning something new. But again, when you think about you said it right, right.
I think we are a speech company and we started with an intent of building a better, understandable world where people can communicate with each other without any friction points. And we thought speech was a big impediment. When you think about understanding speech, people naturally gravitate towards understanding each other’s language. Yes, it’s very important. That’s also a product that we have. But there are so many other things that people discount. Like you’re in a phone call with somebody else and you can’t hear that person well, because either the person is in a noisy place, he has his kids yelling in the background, the TV is turned on. He’s in a place where there is music playing. There are.
He’s probably in a in a metro where people around him are talking like it could be dialects where different people have different dialects, and not everybody understands each other’s dialects really well. Or you just need a high fidelity type of a conversation where like when we are on Zoom, it’s 44kHz. So we understand each other. Like we both are sitting in a studio, but not everybody in a studio all the time. And then it’s language. So there are so many challenges of real time speech. So again, as engineers who got together, we just thought this will be a problem worth solving. One as founders, we we get excited building this building and solving for speech related challenges. But we also realize that hundreds and thousands of enterprises and 8 billion people will all benefit from it. So that’s that’s been the concept of the origin story. I met my co-founders at Stanford. And we got together about five years ago.
And the mission was very simple that, hey, speech is probably the most misunderstood. Taxonomy has the most difficult syntax, has the most complex interface. And if you can build a company which can make two people understand each other very clearly, I think magic can happen, right? We didn’t know who we’ll build it for. We’ll build it for enterprise. We’ll build it for a consumer. We did not know when we started, to be honest, but I think we figured along the way that, hey, I think naturally I come from an enterprise SaaS background. I think naturally this this is something that I can go and take to enterprises. And then along the way, we figured that, hey, why just restrict it to enterprises? Why don’t we open it up to people? And that’s why we started opening up to people. And then ChatGPT happened, and then all the generative AI applications started being built.
And then we figured that, hey, as, as applicable, chances for human to human communications, it can be very relevant for a machine to a human communication as well. Because yes, a machine is in a studio, it is perfect, but it is still trying to communicate with a human who is in an imperfect environment. Right? So we could be relevant for both the worlds and, and, and, and we decided to start opening up our SDK for people to build on. Right. So it’s been a very interesting journey, but for me, it’s personally been a very fulfilling journey as an immigrant, having speech related challenges where I always was loud so that I could be heard and understood. So Sanas for me was, is, was like almost like a referendum that, hey, now people can understand me very clearly irrespective of having fast speech or having a different dialect. Me being comfortable to speak in a different language. I still know using Sanas, I could be understood really well. So yeah.
Dr. Jeremy Weisz: 13:12
Sure.
Sharath Narayana: 13:13
You know.
Dr. Jeremy Weisz: 13:13
But there are so many applications to speech. I’m wondering how did you know early on what problem to focus in on and maybe your first milestone client that you know, because obviously it’s only proven when someone pays you, right?
Sharath Narayana: 13:31
I know, I trust me, we were we were that first two years, we were just a speech lab and we were geeking out. We built almost 14 different algorithms and we were having a lot of fun small team, 25-30 people. And, and then at one of the board meetings, our investors are like, hey, it’s cool. You’re building incredible technology. But at some point you have to start making money. And the first few customers we went to, we went to one of the largest health care companies. We went to one of the largest hospitality chains. We went to one of the largest contact centers and we said, hey, here is the a la carte of speech algorithms we’ve built. What would you pay for? Somebody said, hey, this language thing seems very interesting. What do you think is the latency today? And this was two years, three years ago. And we said it’s between 3 and 5 seconds.
They’re like, no, no, that’s not practical to do. When it’s a second come back to us. And then they’re like, hey, this noise algorithm becomes very interesting because we are in B2B, even when we talk to our colleagues and we have global workforce, noise can be a factor. We should start testing that out. But then all of them came back and said, hey, this whole dialect harmonization solution can be very interesting. One obviously, most of these large enterprises have large teams overseas that, hey, we can understand even internally when we do calls and even externally when we think about customer experience, this can be very useful so that people can be understood really well. And we’ll start paying for it. And we didn’t even know how to price it. We just set out a random number and they said, that’s it, this works. And we’re like, okay. And that’s how we started, right?
Dr. Jeremy Weisz: 15:17
Do you end up pricing like per agent or something? Like how, how do you.
Sharath Narayana: 15:21
Yeah. Per user.
Dr. Jeremy Weisz: 15:21
Per user.
Sharath Narayana: 15:22
Per user. That’s what again, this was in 2023. That’s probably was the most. You the most common way people were charged for a product. We we thought about outcome based.
But then we were like, hey, we’re very early. We don’t know how to own the outcome with a good conversation because different things in a, in a support conversation, the outcome might not be more revenue. It was just, it could just be better experience on sales. it could be more revenue. But when you think about internal communications, I don’t know how to quantify that. So we said, let’s just do it per user and we played around the price. But again, our first big hospitality customer was Wyndham Resorts.
And they came back and said that, hey, it’s funny when we use your product on, on our sales calls, people are spending more time, which means that our dollars earned on that call is higher on our support call. The call duration is shrinking, which means that people are able to understand each other quickly and their problem is resolved and they’re out. Right? And they did it on a good set of people. They started with a small pilot and they kept growing, growing. And then they put everybody across their sales teams, their customer care teams, their customer experience team, their collections team, everybody on soreness. And the results were fantastic. And then we said, okay, is this just because Wyndham, the folks are really nice or is there something else?
Then we went to United HealthCare, very large healthcare account, and they said, hey, our members, we want to we want they’re very important to us. We want to treat them really well. And we want to make sure that both our employees and our vendors can understand each other really well. And they started with a small pilot, and now about 16,000 people within the organization use us. Wow. And and all different use cases, doctors, nurses, contact center agents, salespeople, insurance people, compliance people like the whole plethora of people, right? And then we went to a telco, right?
Dr. Jeremy Weisz: 17:24
What were the features? Just talk about the healthcare for a second. What are the features that stuck out for them? You know, one thing that sticks out to me is like, obviously just the importance that you just went and got feedback from the customers, direct feedback on what they want and what the problem is, what they solved. I’m curious how you described it, like when you described it early on compared to how you describe it now, early on, you’re just kind of like finding the words go. It’s a dialect harmonisation or whatever. Like, yeah. How do you describe it now to someone compared to then? Right. Because then you’re still figuring out what’s the best way to share the, the benefits in a very short period of time.
Sharath Narayana: 18:08
Yeah. I think today we just call ourselves, hey, we have a clarity product, right?
Dr. Jeremy Weisz: 18:13
Clarity product.
Sharath Narayana: 18:13
I think, yeah, if you think clarity in speech is very important for you, it could be anything you do, you could sell, you could support, you could collect money, you could just talk internally between teams. If voice clarity is very important for you, right? And if you if you think that improves productivity, it improves efficiency. If it improves customer satisfaction, then you should use. Right. And you don’t have to use an app. You can use our SDK and build something else on top of it if you want to do it. We we’ve kept it very agnostic on how you want to use it. But coming back to health care that you were talking about.
Different people started with different use cases in healthcare. They first started with just the noise use case. They said. They just said that, hey, our members, most of the times are calling in from a hospital that sometimes under panic, right? They might be in a very noisy environment. They might be rushing to the hospital and they’re making a call. Right? Just checking on coverages, checking on copays. Hey, we just want to make sure that because of that background noise, our employees who are taking these calls, they understand them clearly, at least from our side. We don’t want to keep telling the customer, hey, can you please repeat? I didn’t understand what you were asking, right?
It was not the other way around. It was not customer to an agent. It was more of a employee of United Healthcare to a customer, right? So they started with that as a use case, right. And, and they tested it and they said, hey, our, our, our agents, our employees are now able to understand the customers very clearly. What else can we do? And then I said, hey, we have this dialect harmonizer, call it an accent translation product. If you have a global workforce, then it might just help in improving. And it works two ways, right? And then we’re like, okay, if it works two ways, that’s magic. Like, why can’t we try that? And they started overlaying the accent algorithm on top of the noise algorithm, right?
Dr. Jeremy Weisz: 20:04
So if like someone calls in like me and I’m like, I need to know my benefits and it gets routed to someone overseas and they’re speaking English, but they have an accent, it kind of smooths that out somehow. Is that it?
Sharath Narayana: 20:17
Essentially what it does is in real time, it will still keep the original voice timber intact, but it will sound more local so that you can understand it very clearly. Right? It, it essentially you think about learning. If you think about accent training or like.
Dr. Jeremy Weisz: 20:35
So someone calls from Texas, will the Indian accent sound southern? Not exactly. I’m partially kidding with that. Like, that’d be an interesting combination.
Sharath Narayana: 20:45
No, we’re not trying to match people’s accent. Right. I think we still, like the person knows that if you’re if the call is being routed to Jamaica, the person knows that this is a person speaking from Jamaica. But this the other person can understand the Jamaican accent very clearly. Right. I think if we have to harmonize the rate of speech in real time, if you have to roughen the edges, we just have to make that person sound very clearly. We do all of that. Right? And, and for us, it’s universal accent, right? Like any accent in any accent out, right. Our goal is to enable clarity and that’s what we focus on. Right? And yeah, I think that’s, that’s then they started overlaying accent and then they came and said, okay, this noise and accent is good.
But what we are also realizing is we have, we have a caste system on which the agent is taking a call, right? Or a person on our side is taking a call, right. But a customer is still calling from their cellular network, right. And cellular network is 8kHz. They’re usually on the move when they call us. Right? So you have this constant, the voice is coming in and out.
They just might have a bad headphone. Not everybody will have an AirPod to sound really clearly right. So we sometimes lose a lot of those words that they’re saying. And I don’t know how to fix that. That’s when we figured out building an enhancement algorithm where today we can upsample like we can make a 2G connection sound like 5G, right? And we can recover clipped audio, we can minimize packet loss.
And with every algorithm that we were adding, hearing what our customers were saying, the quality of the conversation between two people started getting better and better and better, right? And, and then we went to a large telco and we told them that, hey, you have 100,000 people in your company. Why do you not want to put on everybody? And they did that. And then they came back and said, hey, why can’t we use it on our network? right. We have 12 million subscribers. Why can’t we use it on our network? And and as of right now, they’re testing it on their network. And the initial results have been very promising. And and now they’re like, hey, we just want to roll it out to everybody. Right. So for me.
Dr. Jeremy Weisz: 23:02
I’m curious on that front for a second, because I get with a company, they and this, this is incredible. Like, right, you have thousands of, they have thousands of employees, right? And you charge per user in that situation. Do you start thinking of like, do they just license the technology? Like, how does that how would that work from like a monetization of 12 million people are using whatever Verizon. It’s like, how do you think about that?
Sharath Narayana: 23:30
I think eventually it will come down to something like that, right? I think see, we are still doing a pilot with a telco use case. It’s new, but the good thing is our customers came and told us they want to do this. And we’re like, yeah, be my guest, test it and let us know how useful it is. And now we have three large telcos from different parts of the world, like somebody in the US, somebody in Canada, somebody in Japan, and very likely somebody in India. And these are some of the largest telcos in those countries using it. And my sense is it will not be a typical per user 20, 25 bucks pricing, because that won’t work. I would love to get paid so much, but I don’t think exactly.
But it will come down to some kind of a licensing fee. Right. And and it will be a large volume, so it will make sense. But ultimately, see, yes, I want to make a lot of money building this business, but the sole purpose of building something like this was, hey, you’ve solved one of the hardest challenges there is where you’re breaking down speech in real time without an intermediate form factor. We do speech to speech, not speech to text or text to speech. Right? And second, we always wanted this to be applicable to 8 billion people, right? And telco can take us to those 8 billion people because almost everybody has a cellular phone, right?
So I’m personally super excited about it because once you have that clarity algorithm on, then you can do so much more. Dr. Jeremy. Right. I look at all fraud that happens today with a lot of senior citizens, right? Like if I can understand true speech, we can detect fraud. We can tell it in real time. If it’s a synthetic call or it’s a human call. Right. I have a I’m a dad of a seven year old. I hear horror stories of kids being groomed to do some shit. Right. And now with if we understand true speech, we can understand that fraudulent grooming behavior on true speech. Right? And then we can notify their parents or notify somebody else. And for all that, you need the scale of a telco. That’s why when telcos approached us, for me as an engineer, I’m like, hey, there are like 100 different intelligence.
Dr. Jeremy Weisz: 25:35
Your mind explodes with all the applications.
Sharath Narayana: 25:38
And I don’t have to build everything myself. I want to open it up so that other people can build all these intelligence stacks because, see, at the end of the day, we want to make the world world a better place, right? For people to live in. Right. And we want it easier for people to communicate with each other. And if soreness has some small role to play into it, I think I can build a generational company. So for me, this is all these use cases are very exciting.
Dr. Jeremy Weisz: 26:03
You know, back to the beginning for a second. What I find interesting is one like your team, you know, maybe built several different models and use cases and algorithms, and then you went to the customer potential customer and you’re like, here, what do you like? What would you pay for? And you got feedback from that. And then you, they started testing it and you rolled it out. I’m curious how, you know, these are big companies. They’re busy people. How do you get in front of some of these people to even present them with, hey, I want your feedback on this.
Sharath Narayana: 26:45
Well, I think C it’s not as easy as I made it sound like. I think the key people who helped us was this whole channel ecosystem that I had built in my previous company, right in when I was building observe, I think outsources like BPOs as the world knows them, right? Have strong channel partnerships because they serve some of the largest brands on the planet, and they listen to them because they support their entire customer experience journey. And very early on, we partnered with Alorica and then with Teleperformance, two of the largest contact centers, and their, their founders, their CEOs became close friends. And they were also fascinated about what we were building.
They were also having that kid in the candy store movement that, hey, this can have so much applications to what we are. Right. And, and when we went to them that, hey, can you take us to a few enterprises that you think who are more friendly, who will give us a listening ear? They were the ones who took us to all these large enterprises, right? I’m always immensely grateful to them. Like Andy from Alorica, Daniel from TP. They had the vision that, hey, I’ll not think about will this cannibalize my revenue? What will AI do to me? They had the foresight that, hey, I think this is great. We can be partners. We can do some kind of a rough share. But more than money, this is a product worth listening to. And they thought about their enterprise customers and they said, now we are bringing true value to them by introducing Sanas.
And that credibility coming in from CEOs of two of the largest BPOs to their end customers got us that meeting. And in that meeting, me being an engineer, yes, I’ve done sales for my companies because I have always co-founded with two other engineers, and one of us had to sell. So I ended up becoming the sales guy. But I’m an engineer at heart, so I just went and told them the truth. Hey, these are, we think, really compelling algorithms, but I don’t know how to make money out of this. You tell me, what will you pay for? And they told us, hey, we’ll pay for this, we’ll pay for this. And I threw out numbers and they said, that sounds reasonable. And I’m like, okay, that’s the price and let’s go. So that’s how we started. And yeah, it’s, it’s been very fulfilling experience.
Dr. Jeremy Weisz: 29:00
Yeah. I mean, it comes back to relationships, right? I mean, people open doors when you have relationships. What did, what did observe.ai what did talk about that? What observe it.
Sharath Narayana: 29:11
Observe was a conversational intelligence company. Right. I think this was early days of machine learning. What we thought was, hey, there’s so much of data around us, right? And if we can synthesize all that data and provide insights to enterprises, we can create a lot of value. And, and we zoned in on before observe. I had, I had my first startup, which I saw sold long ago, and I ended up spending some time in Manila and we were thinking about which use case to build algorithms for. I said, hey, everybody is trying to build it for sales teams. I spent a year in Manila and and there’s this industry called contact centers. They’re like a million people in one city, right? And they keep talking to their end customers that it must be a goldmine of data and nobody mindset, right? So why don’t we build it for that use case?
So we basically build a conversational intelligence system to mine contact center data. And the goal was go to enterprises and say that, hey, that’s a black box for you. You’re recording those calls. You store it for seven years because you have a compliance requirement, but you don’t do anything with it. Why don’t we help you mine that data and provide you insights? And people said, yeah, that’s interesting. But they also came back and said, is that if you just give us insights, it’s like picking a needle in the haystack. What do we do with those insights?
Can you make that insights actionable? That’s when we picked up a few use cases like, okay, I’ll use all these insights to help you monitor compliance. If you are a financial services company or a health care company, if you’re a retail company.
I will use those insights to provide coaching to your people that, hey, you’re representing Saks Fifth Avenue. If somebody is complaining about something not right, you should coach your person in real time and tell them what to say. What is the rebuttal? So we built the world’s first agent assist platform. This was 2018 before copilot became what it is. Right. And then for other industries, we just looked at QA being a very manual function. We said, let’s kind of automate the QA. We went to Bpos and said, hey, you have this random Excel sheet of 20 questions. Those questions have not changed for three, 30 years, right?
Why does somebody have to manually fill those? Yes and nos. We can do it through AI, right? So we built an automated QA assist and this was all in 2018. Right. But I think that gave me access to a lot of relationships, which I could encash it where when I built all these algorithms, I went and told to all these partners that, hey, can you help me take it to market? And today, 18 of the top 20 BPOs are our partners, and they have taken us to over 80 large enterprises. We work with 21 of the fortune 100. I don’t think I could have gone and sold it myself. Right. This partner ecosystem enabled me to go and sell to all these people.
Dr. Jeremy Weisz: 32:09
Did you decide early on you’re going to bootstrap this thing or do you raise money in the beginning?
Sharath Narayana: 32:17
Initially, I was conflicted, right? I was not sure. We thought, we’ll just raise some small seed money and and let’s see what happens. But then I was also a second time founder. My my previous company had done pretty well. Then a lot of people came in and we were when we were trying to build a speech lab, we started figuring out, getting access to some strong researchers was very important. And processing data to build that data mode was very important. And all that consumed money because compute was not.
Dr. Jeremy Weisz: 32:48
Expensive.
Sharath Narayana: 32:49
Cheap as it is today. Right. Five years ago. So that’s when I decided, okay, let’s let’s found a find an investor who is patient, like look at more patient capital because the today everybody understands building a lab, right? People are like, hey, you can spend $200 million, no questions asked, but build something interesting, right? But back in 2021, it was not many people understood that, hey, you need to two years you will see nothing and then you will probably see something. And I have to be truthful and insight was that patient investor. We spoke to a bunch of people and they felt like I felt talking to them that, hey, they’ll they’ll play the long term game that a a large fund. And they don’t seem affected by somebody else growing quickly.
And and they were patient for two full years. I think we tested their patients after two years where we were still not never talking about a commercialization model. That’s when they said, hey, I think you have to go out and figure out how to make money. That’s why we decided to raise. But I’ve not raised too much. If you think about companies of our size, $60 million of revenue, that the growth at which we’ve had, people have raised hundreds of millions of dollars, we have so far raised about 100 million across four and a half years. We have profitable in our core, we have 96% gross margin. So we’ve tried being responsible and be capital efficient in our journey. But but yeah.
Dr. Jeremy Weisz: 34:19
It’s pretty incredible what you’ve done. I’m wondering about hiring. You’ve hired a lot of people in your day and, you know, maybe some methodologies around how you decide when you’re hiring people.
Sharath Narayana: 34:35
Again, across my three companies, Talk to me. I’ll tell you, there’s no silver bullet to that. We all make mistakes, and I’ll be the first to admit that. But I think what we’ve tried doing is we’ve tried hiring people who at who, at least we feel during the interview process, are genuinely excited about why we are building science. Right. The overall mission statement, right. And it was natural that the first I think the 44 people that we had hired were all immigrants. Right? They came from somewhere. They had genuine challenges with language, accent.
Dr. Jeremy Weisz: 35:14
And they get the pain points.
Sharath Narayana: 35:15
Being understood right. I think it felt natural. Right. So and then obviously now we have people from all over the world who work for us. But I think that initial core, they connected to what we were building. And, and then afterwards, all the other people we’ve hired, they came and saw the people who were in the company. If you look at our researchers, the initial eight researchers we hired, they’re all still there, right? There have been every large company on this planet have tried to poach them. Yes, I pay them well, but I don’t pay them like a meta or an OpenAI. But they’ve not gone because they have kind of connected to that mission. And they feel that, hey, you’re yes, you’re making money, but at the same time, you’re serving a greater good, right?
And I think it becomes important for a company to have a soul. I think that has helped me hire people a little more effectively. But I make mistakes every month. But what I’ve decided is, hey, if I make a mistake, I would share it with that person sooner than later and see if something there could be turned around. Otherwise, we just mutually part ways. And I make a statement in every all hands that I do every month that hey, please be in Sanas because you want to, Not because you have to. Ever. You feel disconnected. Just let me know. I’ll give you three months. I’ll connect to you to more people in my network. And let’s have a very respectful exit. And that’s the only way to build. But still, I make mistakes.
I would just say that.
Dr. Jeremy Weisz: 36:54
Yeah, I like how you connect it back to really the vision and the mission, and people are on board with what you’re trying to build. You know, Charlotte, I know you went through Y Combinator. I’d love for you to talk about some of the mentors colleagues. It could be colleagues too. And some of the lessons you learned. It could be you could start at YC and maybe just overall, who are some of the people in your universe that have helped in your business journey?
Sharath Narayana: 37:22
No, no. For me, see, my first startup I built in India, I came to the US because we got selected to YC and it was my first time building here. I had no network, no connections. So YC became my safe space where I had 50 other people like me. YC batches were still very small then.
5060 people like me who are probably as clueless as I am. But then you had folks like Michael Siebel, Sam Altman, right, who built generational companies today just trying to be helpful. Like all YC did was it created a safe space to founders. And they you go and ask for help. They’ll help you. If you don’t ask for help, they’ll leave you on your own.
So it never felt like it’s like a micromanaging investor trying to poke holes to everything you’re doing. It just felt like an ecosystem where you could be yourself, you could talk to like minded people. And just like, I think probably the only time I’ve gone and told to an investor that I fucked up was at YC, right? Because it created that safe space to share it. Right. But I think that community, trust me, is probably the best community that we’ve had. I think all our initial customers and Observe, or at least initial people who try decided to give us access to data, test and tell us where we went wrong. We’re all XYZ founders. And not every day you can write an email to the founder of Coinbase and Airbnb and they’ll actually respond, right?
Dr. Jeremy Weisz: 38:59
So who are some of the interesting companies in like your, I don’t know, in your class or right around there?
Sharath Narayana: 39:08
I think there were a lot of them. I think this was Winter 18 when we came in. We we were one of the cool companies in that batch. We had. Even at that time, we had a robotics company.
We had a furniture company called Orange Woodworks. We had a legal firm. Like today we hear about Harvey, but Justin Kan had started a legal company. He was probably too early, but very inspirational at that time. Who were up here. So a lot of them very, very inspiring. I think today when I look back. Most of those companies are still alive, which is funny because it’s been almost eight years since I graduated out of YC to still find most of them alive, including observed. Being alive is fascinating because I think he kept us all grounded and, and that community still exists, right? I think some of them, some of the founders might have gone out and done other companies, including me, but most of those companies were built to scale. A lot of them did get acquired as well. But it was it was a very fun experience.
Dr. Jeremy Weisz: 40:22
One last question. I first of all, I appreciate you sharing your journey with us and people. Check out Sanas.ai to learn more. And I guess just any resources, like are there any favorite books that you’ve learned from maybe some favorite podcasts you learned from anything that have helped you as far as that goes?
Sharath Narayana: 40:45
I think those inspirations have evolved over time. But I think 20VC Podcast is something that every founder should watch. I think even the latest one from Adam from Applovin. I think it’s probably one of the most inspirational podcasts I have watched in the recent times. There are some incredible books where if you’re building as a founder like Netflix story, No Rules Rules.
Basically it tells you about how to build that winning culture, how to trust your people, but also expect everybody to give your best. If you are low, go and read Shoe Dog. I think nobody has had more push to the wall situations than Phil Knight has had. And. And every time I read it, I.
And I’ve probably read it ten times, and every time I read it, I walk away with a very different meaning. And then I think, why has a startup school? You don’t have to be a YC alum to be have access to startup school. It’s now available everywhere. I think there’s like a gold mine of information there.
Stanford very recently has made most of their courses public. So you don’t have to go to Stanford to understand AI and machine learning. You can just read it. I think like how we are trying to democratize speech with Sanas. I think information has been democratized in the AI era, right? I don’t think today people can say, hey, I’m not in the Silicon Valley, so I can’t build a great company. You can build a great company from anywhere you are because the information has become so democratized. So seek out information. And I think generally, even me, I try my best to spend at least three four hours every week, However busy I am just to randomly talk to founders who might have zero relevance to Sanas or even to my investment fund just to help. And most founders, I believe, especially the ones who are successful, always want to give back to the community. So I always tell people the the most important question that you never get answered is the one that you never asked. Right. Just ask for help.
Dr. Jeremy Weisz: 42:55
I want to be the first one to thank you. Everyone check out Sanas.ai. It’s just incredible what you’ve done. And you know, with Sanas and in previously in your journey. So thanks for sharing that and the lessons and everyone can check out more episodes of the podcast as well. Thanks so much.
Sharath Narayana: 43:11
Thank you so much.
