June 13, 2023

Voice AI: Breaking Barriers - Sourabh Gupta | BCL #292

Sourabh Gupta is the Co-Founder and CEO of Skit.ai, a global conversational Voice AI start-up. Sensing the need for better engagement between consumers and brands, he co-founded the company in 2016. Presently, Skit.ai comprises a team of 200+ employees and caters to over 100 clients in markets such as  US and India. Sourabh has garnered numerous prestigious awards for himself and Skit.ai, including being featured in Forbes 30u30.

Sourabh: Ineed to start a company. And I think from that moment on, I couldn't thinkabout anything else. Sometimes you just think about it and you have to do it.Yeah.

Julian:hey everyone. Thank you so much for joining the podcast. Today we have SourabhGupta co-founder and CEO of Skit.Ai, a global conversation, voice AI startup,sort of, it's so hap exciting to have you on the show.

Not only are, am Iinteresting and interested in, in particular, where the AI space is going,especially, with his voice and, and how that is getting integrated in a lot ofbusiness and companies. But in particular, looking at what you're doing at.Skit, which is really interesting as you're kind of, you're not changing kindof an industry that we're super familiar with but you're kind of upgrading itspotential.

And so I'm curiousto think about how your tools are doing that for call centers, for collectionagencies. Before, before we get into all that good stuff,  

what were you doingbefore you started the company?  

Sourabh: Istarted this in my final semester of my undergrad. So I found my co-founder andwe both started straight out of college.

So, Pretty muchstudying before that.  

Julian:It's, it's crazy to think about going straight from college to startup and, andlike what in particular, made you crazy enough to actually pursue the ideawithout any previous experience. What, what was so compelling about, theproblem that you were solving and, and what was the inspiration essentiallythat, that led you this far?

Sourabh: Ithink at some point I felt that I need to start a company. And I think fromthat moment on, I couldn't think about anything else. Sometimes you just thinkabout it and you have to do it. Yeah. I think it was that more than anythingelse. And that led me to it. It was almost like the opposite was crazy for me.

It was not crazyfor me to start a company, but. To not start it and find a hying job aftercollege and just like settling into that. That was sort of the crazy part,so,  

Julian:yeah. Yeah. Well, thinking about like where you were looking for ideas at thetime, and a lot of founders have the same experience.

They feel like,their journey is to lead a company, a product. They, they're so compelled byit, and then they start to uncover ideas through conversations or through trialand error.  

What in particularhelped you kind of identify, what you wanted to kind of build into?  

Sourabh:Yeah, so I think we, we took a couple of years to figure out what to do.

First of all I thinkwe told ourselves this is a very long journey. Um Right. We are in here for thenext 10, 15 years, and if it's a 10, 15 year decision, it doesn't matter. Evenif I spent one year or two years figuring out what to do. Right. So I thinklong term thinking sort of helped us with that.

And, and when wewere doing that, we were trying to, come with all sorts of frameworks, right.You make you make a spreadsheet, Hey, these are the industries you wanna goafter, and so on and so forth. But I think one thing that stuck with us was, werealized we are very young, we are good with technology.

Yeah. So we wantedto work at the edge of technology and we felt that is where it's gonna bereally hard for anyone more experienced to compete with us. And we were just exploringa lot of AI at that time. Blockchain ai, all of that. Yeah. And somehow we, gotinto voice ai.

It just made senseto us. At the same time we were seeing another trend happening in India. So,2016 we were chatting earlier about that, how in India most of the populationnever even experienced computers or laptops. Yeah. Yeah. So they, they, thatwas a mobile first generation.

And this, thesepeople were coming online a few million a week. Yeah. And we felt like thesepeople, when they come online, their needs are gonna be very different from,people who are living in the tier one. Most of these were tier to tier threepeople. So we sort of were trying to merge, the edge of technology with thisnew set of people who are coming online.

And that led us toY A I and that led us to. So four centers and we ended up, where we aretoday.  

Julian:Yeah. And,  

and why voice AI inparticular? What, what about its impact or its usability or was another companydoing something that was innovative at the time? What about voice AI inparticular?

Sourabh:So I remember, so you know, when these people are coming online, I think wewere thinking, Hey, how do we know how to, for them, the best way to do that isgo and meet them. Go and talk to them. Yeah. And I remember going in the tiertwo, tier three areas of the of, of India and talking to these people. One ofthese visits we met a villager, a farmer.

Who, who couldn'tspeak English. So we used to carry a translator with us. Yeah. India hasofficial languages. I only speak three, including English. Yeah. So I, Iremember I took a translator with us and the, the farmer told us he receivesSMS from banks in English. He doesn't know what it, what they mean.

Yeah. But he knowfrom bank because they're digits then he travels 10 kilometers away on abicycle to get it translated from a bank manager. In the process, he loses hisdaily wage. So a large part of, people don't know how to read or write, butthey can speak their local language. And none of the local languagetechnologies in voice existed in India.

So we were thefirst one that built speech to text, text to speech across Indian languages,which you could use in like real life, yeah. Scenarios.  

Julian:Yeah, it's so fascinating thinking about, that whole experience and, and kindof the systemic and operational and kind of inefficiencies that are just kindof present in a lot of just technologies that exist without thinking aboutsolutions like that, thinking about the impact of just that idea in particular,did that kind of, rejuvenate or, or I guess.

Honestly, just pourgasoline on the fire of what potential you could do as founders, but also whereyou could take this technology. And, and where did it go from there? How, howdid it get from that example to call centers?

Sourabh:Yeah, so we started doing this, so we knew there was a gap between, financialinstitutions and these tier two, tier three people living there who can'treally drive properly, especially in English.

So we startedfiguring out how to bridge this gap. And we realized, the, the one of thechannels is call center. So we had to start somewhere. We said Call call centeris where huge amount of volume is there. People are trying to communicate.Processes are broken, and voice second, play a very, good role.

We saw anopportunity there. So we started that. Yeah, we started that in local languagesin India and the moment we started doing that, we started seeing demand forEnglish in India. So we're talking about India, English in the Indian accentand so on. India is the second most. India has the second most populated from aEnglish speaking perspective after English.

Yeah. Right. So, sowe had that and we started building that. So at some point we were supportingonly local languages. Then we started supporting English and local languages.And as we were doing that, we realized the same problem exists in the US and wewere seeing, demand with that. And that made us, that made me move here.

So we started withthat and then we sort of kept, moving forward where the  

Julian:demand is. Yeah, yeah. Obviously we're, we're about to just stumble into someamazing territory, not only talking about, your transition in tr in intodifferent markets and what that experience, I'm sure brought you in in thatchange, but also thinking about the overall nature and, and the The reliance wehave and that everybody has on call centers and what that actually meansbecause there's a bunch of technologies in between, the ability to need arequest and, and get it alleviated, confirm, solve whatever.

But we still alllean towards voice and, and there's reasons for that and I'm curious to hearyour thoughts. But before we get into all that, Just taking a step back interms of AI and the growth of machine learning and when you two started tounderstand and build and iterate in terms of, learning the technology,  

what was like themajor milestone or development in AI that kind of helped propel, more of thesophisticated things that it's able to do even today?

Sourabh:Okay. I think like at some point this is 2015, 1415, yeah. Couple of paperscame out in deep learning. Yeah. I think that was the first time, you as a 23,24 year old could use, the concept of deep learning and build something thatyou can use in production. Building a speech to text before that was somethingyou would leave to larger companies like Google and Meta, and you would expectthem to have the resources to pull it off.

But that learningsort of made it easier for any startup to which I think like in some sense,elements are doing as well in different way today and I'll, I'll get into thatas well. So I think that enabled us to build something that we could not or wehad to rely on, these bigger companies to start doing in terms of speech, text,text, speech, and local languages.

Using that andwilling on top of it, we've been able to create a voice assistant, which is,yeah. And our customers tell us this, like, your voice assistant is at least asgood as our average  

Julian:agent. Yeah, yeah.  

Sourabh:Now with lms, coming to picture, like after deep learning, this is the next bigthing. For now, after this, like our internal goal is we believe we can build avoice assistant, which is better than the best agent.

Yeah. Because it'smore conversational. So I think these kind of things are like elevating orupgrading the experience that one can have with their head.  

Julian:Yeah. And, and, jumping into,  

what I just kind ofalluded to in terms of the reliance on call centers in particular, like, youknow what, why are we so reliant on that experience, having somebody kind ofwalk us through, whether it's a ticket or an issue, or even just informationabout a product or service.

Why, why, why arewe still relying on that when there's. There's chat bots, there's FAQs, there'seven AI and text. Why, why do we, why do we lean like that as humans?  

Sourabh:Yeah. I think like there is an option for other channels also. I mean, thereare scenarios where you use chat. There are scenarios where you, sort of useemail, et cetera.

But I feel like byand large, if you face an issue, You want to call, you want to get it resolveas soon as possible. Yeah. And the trust factor you have with, you don't knowwhen you'll get a response to an email. Yeah. You don't know any response to achat. But you know, if you call a call center, there will be someone on theother side.

It's unfortunatethe wait times are unfortunate. Yeah. Right. But, but, but you do know that youcan call a number and there will be someone who picks up on the other side. AndI think that comfort of predictability is sort of something that  

Julian:drives us. Yeah. Yeah. It's like we have, it's almost like we have somebody onour side, and, and it's, rather than it's us as an advocate. It's us and ouradvocate and having that relationship. And, one thing I think about, with voiceAI technology, and, and I'm, I'm curious to ask you this question is, doyou,  

do you think about,the, the voice ai experience and trying to get it as, Close to being authenticas possible in terms of like a human experience, or do you think about, how tocreate a unique kind of standardized voice AI experience like now we have withchat bots and now we have with all the other things, first it was like theautomated voice dial message.

Everybody wasthrown off by it, but then it's like, oh, press one for this. Do you thinkabout making it as close to human as possible, or standardizing that uniqueexperience?  

Sourabh:Yeah. Ju what we do internally is we come from user backwards. So we wannadesign a best experience for the end user.

Yeah. Which notnecessarily is humanlike or not. So if we end up close to Humanlike, that willbe a functional, what the user wants more than trying to build something, whichis human, right? Yeah. Right now the state of the technology, we believe, likewe, we tell people proactively. This we are calling on a recorded line sopeople don't end up getting confused.

And I think it'simportant from an ethics point of view as well, right? That people know they'retalking to a human or a bot. So you know yet to see what the ideal sort ofproduct looks like. Yeah. But I, I don't think it's gonna be human. Like I, Idon't think, like that's the, yeah. Best way to think about it.

If it ends up likethat, then perfect. But yeah, I think you can use AI in so many ways, which is,it'll be, in some cases it'll be better than human. In some cases it'll beworse. But in itself is what I, we personally believe.  

Julian:Yeah, yeah, yeah. Thinking about obviously the company and, and seeing how it'sgrown, what's been exciting, not only to, to kind of you jump into a new marketand  

what has beeninteresting about that transitioning, growing from, India and that ecosystemnow into the US and what's, what's the same, what's different?

Sourabh: Ithink like we, we almost started another company. That's how we look at it. I,I remember when I moved to New York two years ago, I, I did not know anybody inNew York. So there was zero people in you. And somehow the, the idea was, Hey,we need to start building a business. Yeah. How you wish business when youdon't have one, you don't know anyone at all.

Right. So it wasalmost, journey from scratch, hiring the first employees, getting the firstcustomers. There is some overlap we have within India. I mean, we have a broadsense of the ecosystem in terms of mm-hmm. What product we want to build, whatcustomers expect, et cetera. But we've pretty much had a new journey from start.

So we've done twostartups is what we say. And I think that itself has been very exciting in, ina sense that we've made a lot of mistakes. Right. And we sort of learned fromthem, but when we were doing in the US we already had so many learnings of India.So I think a lot of things happened very quickly and now we are already ontrack to, get so many customers.

Julian:Yeah, yeah. Give us a little bit insight, what's particularly in terms ofnumbers, what's been exciting about the growth you've seen up to this pointwhen and what's particularly exciting about the next milestone that you'relooking forward to?  

Sourabh:So I think in terms of the numbers, like we, we have a few dozen customers.

In India, US, andSoutheast Asia. Mm-hmm. And I think we are doing, we have around 200 employeesagain across these countries. Yeah. I think we're doing somewhere around 10million calls a month. Wow. We, we sort of handle, we support some of thelargest banks, insurance companies, telecom players across countries.

Yeah. If, if you goto India today, And you call a large enterprise and you hear a more on theother side, like, high probability is powered by us. Yeah. So, that's sort ofthe, the scale and the traction. We've raised obviously a few rounds in ourjourney. Yeah. And and, and the next milestone is, just like I said, we want tobuild the best product that we want to make our customers happy.

We want to deliverover value. So that's sort of what we're focusing on.  

Julian:Yeah. Yeah. What, what's been  

in regards to thefundraising process how, how has, obviously the landscape changed and have youchanged anything about how you operate as a company, being that startups kindof face this new challenge, having to operate really efficiently and reallylean and, and, and be hyper productive.

Sourabh:Yeah, I think like, the lens has stayed strong growth at any cost to, so likeit has reflected in doing a lot of reviews of gross margins and, expenses andthings like that. Sales and marketing efficiency and so on, so forth, which Ithink sort of took a backseat in the last couple of years.

So, we've moved orfocused on these things, but again, like that doesn't mean that we are notfocusing on growth. So we still focusing on growth. We still want to getcustomers, but just not at any cost. Like we are questioning the right channelsand which is I think, very thing to do.

Julian:Yeah. Yeah.

It's exciting thinking about, kind of AI andits potential and possibilities, but I, I obviously read an article written byyourself thinking about the, the biases with ai, and I'd love to kind of, foryou to share your ideas in terms of how. Not only we can kind of change ourrelationship and how we think of ai, but maybe even something proactive that, whatcan we put in place to, to help, I guess, I guess, not, not police, but I guessutilize the tool in the right direction.

Cuz like any tool,it can be used to harm and, and for good. But curious to hear your thoughts onthe biases and, and what we should do to overcome them.  

Sourabh: Ithink today. Interesting question. First step is acknowledging that yoursystems might be biased. Like even if they're not, like, I think the rightframework is to think they are.

Mm-hmm. And thefirst problem to solve is like, how do you measure bias in the company or inyour products, in your, I think that's a really hard problem. There are notools or standards in the world. Like, it's not like revenue that, you have anumber and you can say, Hey, this is how biased I am.

But I think likefor everyone need to understand, like lms, if they're trained on data,available on the web, like yeah, by definition they are biased. And whateverproduct that you have, whatever you are building, you have to take care ofcertain things that it's not, bias based on age, gender, religion, et cetera,et cetera.

Right, right,right. So what what we are doing is we're trying to. Sort of build a frameworkinternally of how do we measure it? And then we are trying to work on, after measuring,we're trying to work on how do we design an AI architecture which reduces it,if not eliminates it a hundred percent.

But again, like Isaid, it's, it's an active area of work. And I think everyone needs to startforcing on that.  

Julian:Yeah. Yeah. And what, what are some ways that you, what are some things thatyou've put in place to kind of alleviate or acknowledge or even just I guesssignal when those biases are, are coming about?

Sourabh:Yeah, I think like one of the things we've done is based on if in our callcalls, we have access to data like gender or name. Mm-hmm. And we are usingopen source technology like LLMs. We should not be having conversationsdifferently with people. Yeah. Just because you are Julian and I'm sorry, shouldnot change the response the voice consent is giving to you or me.

So that is anexample of, what we have sort of done. Yeah.

Julian:Yeah. So fascinating thinking about where that leads to and I'm, I'm curious tothink about, Taking a step back, looking at the company,  

whether it'sexternal or internal, what do you view, as some of the biggest risks that thecompany faces today?

Sourabh:Yeah, I think like with advancements in AI that we are seeing with Netherlandscoming in, like it's, it's well known real that we are heading towards theworld of what we are doing. Yeah. Right. So the question of Will Royceassistance become a norm across the world is not a question anymore. I thinkthe answer there is yes.

The question is whowill do it more than will it happen? And that sort of tells us that the biggestrisk for us is execution risk. And what I mean by that is like, hey, are wefocusing on the right market? Are we, do we have the right product strategy? Arewe keeping our customers happy? Are we taking care of churn?

Are we taking careof our costs while we're building? And so on and so forth. So I think like, webelieve execution risk is the biggest, this is big enough risk in itself. SoI'm not trying to say that there is no risk. But I, I think like we are in agood place right now.  

Julian:And, and

if everything goes well, what do you, what doyou think, what, what would you say is a long-term vision for the company?

Sourabh:So I think we've always had a vision of building a unique, differentiated andsophisticated voice AI technology company. And what I mean by that is like, ithas to be unique. It has to be differentiated and has to be sophisticated. Sothat's sort of the backend, what we are doing today. If I'm working in, thedebt recovery market automating bots, et cetera, it's, it's short term.

Call centers isshort term. We have the vision to Beta voice, air company, and that's what weare sort of moving towards. Working in call centers gives us access totremendous model data which we can use to keep building our systems. Voice, airtechnology for the future and what we do strategically every 18 months changes.

And it'll keepchanging as long as but we will always be building voice ai.  

Julian:I, I love this next section I'll call my founder f faq. So I'm gonna hit youwith some RapidFire questions and we'll see where we get. So, fir, firstquestion I always like to open it up with is what's particularly hard aboutyour job day-to-day?

Sourabh:Wow. Okay. I think like context switching. Yeah. 30 minutes you're doing a techreview and then you're doing a product review, and then you're doing salesreview. Asking the right questions and getting in that zone is sometimes, youknow, hard.  

Julian:Yeah, yeah. Thinking about, time in particular, I think a lot of foundersobsess on, on where they spend time, having more time needing more time.

What in particularwould you say something that you spend less time doing that you would like tospend more and a lot of time doing that you would like to spend less?  

Sourabh:Okay. That's a good question. I think like, maybe the second one. So what I doless, which I should do more, is like communication with the team.

Mm-hmm. We have 200people I'm here. A lot of them are in Bangalore or I media. So justcommunicating with them one on one, even one too many. I think I should startspending more time doing that. And what I should stop doing is I think like, Istart going deeper into, different functions like in sales or in success, etcetera. I should just leave it to the leaders and take a step back.  

Julian:yeah. It's such a founder mentality though, to, to try everything beforeanyone's done, that job function to be able to understand. Right. I feel like.It's like time and time again. I mean, I was reading the hard thing about hardthings the other day and it was like, that was just like drilled into a lot ofwhat he was talking about.

And you see the,the reasons why, how it's successful. Thinking about obviously just like, a lotof founders who have co-founded relationships talk about challenges, but alsohow you kind of divvy up responsibilities, go through kind of hardships, how tomaintain a strong relationship.

What are some thingsthat have helped you and your co-founder kind of, have a strong relationshipbut also keep each other accountable, which is, what we all hope to have.  

Sourabh:Yeah, I think the relationship with me and my co-founder from day one was whilewe both come from technology background, he was the tech guy.

Yeah. And I wasdoing everything else. I think the boundary was very clear that, hey, I, Itrust you with technology and from his point of view, I trust you witheverything else. Yeah. So that's sort of something we maintained for, for awhile. Yeah. Until recently we got a VP engineering and that sort of led us toa, point where we were questioning hero, what is the best place for myco-founder to start spending time on?

And he moved intoan operations role. I think one of the things we realized, especially given I'mhere, he's in India, very regular communication is very important. Yeah. So oneof the things we did was set up a regular one-on-one meeting on the calendar.Even though you are seeing each other in different review meetings, I thinkit's very important to do that.

So I think that issomething that has helped us quite a lot. And in those meetings we discusseverything, including, some of the question, what did you do in the last week?You said you'll do this and.  

Julian:Yeah. Yeah.  

Thinking about,other companies out there I always like to ask founders this, what, what aresome companies doing, some really cool things, either from a technologystandpoint or from a creative kind of user experience standpoint that you're,that you're a fan of and that you kind of hope to emulate with the companyyou're running right now.

Sourabh:It's like two companies are coming to my mind. I think one is obviously we haveto talk about open ai. Right? I mean, it's just so cool what they've done andthey were under the radar for so long. Yeah. And all of a sudden, it's, it'sjust blown up. They've created something amazing And Sam Altman is like nowgoing all over the place.

Yeah. Residents andprime ministers of different, I think like that's really cool as a founder tohaving created something that everyone the world is acknowledging and, everyleader in the world wants to meet you. Yeah. At this point. So I think that'svery inspiring. I think the other something that we use today to today isRevix.

Mm-hmm. Usingrecent, it's, it's, it's so smooth. It's, for the longest I was trying to figureout what to do with my invoices, how do we manage that accounting, et cetera.And I think the user experience they have created on both mobile and, and web.Yeah. For both the finance team and individuals.

And you can get acard whenever, I don't know if you've used. Bricks. But it, the user experienceis beautiful.  

Julian:Yeah. I don't use them personally, but I'm a big fan of them because they hirea bunch of developers from South America. And I, I run a, run an agency that,that that's where our main source of candidates are.

But they've justbuilt such an impressive company and are such a smooth experience. They'vemaintained such a strong brand. Throughout the years that, I mean, I think theylook the same. It's like still sleek, still, kind of cutting edge, but, kind ofrigid and financial. From now from I think, I don't know when they started it,it's pretty impressive how to, maintain that consistent image.

Thinking about, thetechnology in particular, voice ai, obviously you're, you're working on, whatskit is, in the direction it's going. But what would you hope to, or maybe ththis type of technology, what are you interested in its application in the, inthe grander scheme of things, whether it's skit or whether it's somebody else?

What, what's inparticular interesting in terms of where you can apply it?  

Sourabh:Yeah, I think like one of the things which I feel will get upgraded. Is havinga personal assistant. Yeah. And, and when I think about Alexa or Citi, I mean,I have two Alexas at home, for example, in my living room.

But you can onlyget so much done. Yeah. It can help you with alarm, with the reminders. Maybemy grocery list and you'll screw that up well, a couple of times, but I thinkthe, the scope of improving that is so significant. Right now, and I don't knowif you've seen the movie Heard.

Julian:Yeah. Yeah. I, I've never, but I've heard of it. Yeah. Yeah. Oh yeah. So  

Sourabh: Ithink like, I mean, that is obviously a very extreme angle to it, which ishaving a, almost like a girlfriend. Yeah. But I think these personal assistancecan get so much more done. With S uh uh, and even if you just plug in GT fourthere, then that is an area which I'm personally very excited about.

Julian:Yeah. Yeah.  

It's so exciting tothinking about the different tools and plugins, the expansion of, of all thisAI technologies. And it's, it, it, it's, are you, are you, what side are youon? Are you on the side that, there's gonna be this terminator style apocalypsethat's gonna come? Or is it just gonna continue being a stronger tool?

And, and, takingaway a lot of the, inefficient work that I think we do manually to really onlyhave us focus on important things or, or things that are timely. Which side areyou on?  

Sourabh:Yeah. So I think like my framework is like in these kind of things I rely on,you know what, what, like the really smart people were working in the edge aresort of thing.

A lot of them areworried that sort of worries me as well. So I think like regulations areimportant. Yeah. In some shape or form. But like if people like Long Musk aresitting there worried about the information they can have, I think likeeveryone should be  

Julian:Yeah. To extent concerned.  

Sourabh:Yeah. I dunno if you'd ever end up in the Terminator kind of situation becauseI think regulations will kind taking care of someone, but it is sort of scary.

Yeah. Like, maybenot that, but with open AI and pretty things you can do with just, you cancreate a greedy of a president. Talking about something and it feels real.Yeah. And the way that can be used or misused is Yeah. Is just veryharmful.  

Julian:Yeah, it, it is cool though to think about, its adoption into the, the speed atwhich technology gets developed.

Like, there's beenso much, significant advancements, not only just thinking about whatmicroservices did, but also now implementing AI and all these, almost likehyper-productive, tools and engines. Discuss, I don't, I don't know if you haveany reflection on this, but you know, from who you talk to or who you meet, oreven, Technology, how much different is it building in today's world than evenlike 10 years ago or, or even five years ago in how available, building asophisticated product is.

Sourabh:Yeah, I think like when we started building has always been, I think, equallyhard. What you can build has. Changed, if that makes sense. Mm.  

Julian:Okay. Yeah.  

Sourabh:So, like that, like I, we could not, you could not build a board as good as abest agent earlier. You could only build an average board. But building anaverage board six years ago took equal effort as building the best agent boardtoday.

So I think likethis just enables us to upgrade the products when it's always been equallyhard.  

Julian:Yeah. No, that's a a hundred percent no, that's a great perspective. I love, Ilove that perspective. Definitely something different that, that I haven'theard yet. I always like to ask founders this, cuz I love, what you extract andwhat you ingest knowledge and where you ingest knowledge from, whether it'searly in your career now,  

what books orpeople have had a lasting impact on you as a founder?

Sourabh: Ithink we did talk about on Musk, so I know, I know it's a cliche answer for anyentrepreneur, but I remember like back in the days, was very young in my lateteens, et cetera, I was reading a lot about him. I read his back, him lookingforward to his next one. And the guy who wrote jobs, it's coming in September.

So I'm lookingforward to that one. But I read the Ashley Vance one. I think that was veryinspiring. But beyond, like cliche, like last year I read might be mixing ofthe name, but Steven Schwarzman. Steven Schwarzman, the Blackstone founder.Mm-hmm. He lives here. Right here in the upper side. Yeah.

Yeah. I think I, Iread his autobiography. That was very inspiring. And, a finance guy, somethings were I, I picked up, he, he rarely ran being a finance guy, York.  


Sourabh:Things very inspiring, the way build Blackstone and  

Julian:Yeah. Thing right here. Yeah. My I'm still trying to get into the SchwartzmanScholar program, so if he sees this he, he, he's gotta put me in throughadmissions push my application up.

No, people like himare extremely brilliant and, and and you learn so much. And also they own somuch of, so many assets of the world, them, along with Vanguard and but thestrategic be how you do that and how you put. Systems in place, how you buildpartnerships and how things feed into an ecosystem.

Once that startsto, once that flywheel starts going it seems like endless at, at some point.And so I guess my question to you is, what's that moment at least, within theshort term that gets that flywheel going to, to, lead you to that nextmilestone? What, what would you say that is?

Sourabh:Yeah, I mean, I think like we have some goals with respect to building, ai Ithink we have some milestones for the next year, for example. I think at somepoint, like I said, it goes both at any cost and we are growing, et cetera. Andthen, when the market, the macro changed, we learn things and then we decided,hey, you know what?

We have to movefrom there into growth at, reasonable cost, et cetera. Yeah. And I think gettinginto some of those cost structures can take time. Yeah. At the scale of 200people, you can't just like overnight say that, Hey, I'm gonna, improve mygross margins by 50% overnight and, and so on, so forth.

So I think we, wedo have some of those goals in our mind that, we, we are sort of going forwardwith. And, but like I said, like the re recently, one of the things we aregetting very excited about is building agent, which is. Building on what, whichis better than the best agent. I think that is something which is superexciting goal for us internally.

We have team ofmachine learning engineers that are just working on that day in and day out.Yeah. So hopefully, but you know, the changing milestones problems. So themoment you are close to one, you know they have five goals that you have.  

Julian:Exactly. Exactly. Well, sort of, I know we're close to the end of the show, soyou know, I always wanna make sure we didn't leave anything on the table.

Is there anyquestion I didn't ask you that I should have?

Sourabh: Ithink we are good.  

Julian:Well, Sourabh thank you for being such a, such a great guest and, and honestly,not only diving into your background, your experience, but how you view kind ofnot only building your company, but you know, the building landscape, thefunding landscape, and, and really giving, giving a lot of anecdotes that Ithink companies can, use to, to their business and, and how to think and, andbuild and strategize and even who to look at to do so.

So, last little bitis where can we find you? Where can we be a fan of you as a founder? Give usyour LinkedIn, your Twitter. Wherever you're most active, where, people on thepodcast can get in touch and be a fan.  

Sourabh:Yeah, I think LinkedIn and Twitter those are two platforms where I am availableall the time and personalize this private.

So I won't sayInstagram or anything else.  

Julian:Love it. Love it. So it's been such a pleasure walking, the audience throughyour journey, but seeing where you are, where the company is and really justgetting excited about its potential. So, I hope you enjoyed yourself and thankyou again for being on the show today.

Sourabh:Thank you, Julian, for inviting me. Thank you.

Julian:Of course.  

Other interesting podcasts