Praveen Seshadri: Building AppSheet - Five Pivots to a Google Acquisition and the AI revolution

In this episode, we sit down with Praveen Seshadri, former Microsoft executive and co-founder and CEO of Thunk.AI. He previously co-founded AppSheet, which was acquired by Google. 

Praveen shares his journey from leaving Microsoft at age 40 to navigating five pivots over two years before discovering product-market fit with AppSheet. He discusses how the company pioneered no-code and product-led growth, the importance of customer-led development, and the lessons learned from building and scaling a startup.

The conversation also explores his latest venture, Thunk.ai, where he is focused on using generative AI to automate business workflows. He shares insights on the current state of AI, the challenges of reliability, and why AI has the potential to fundamentally reshape how work gets done.

In this episode, Praveen talks about:

- His transition from Microsoft to entrepreneurship
- The journey from multiple failed ideas to building AppSheet
- The rise of no-code and product-led growth
- Key lessons on fundraising, team building, and customer focus
- The opportunities and limitations of AI in business today

Whether you’re an entrepreneur, technologist, or builder, this episode offers a candid look at what it really takes to build, pivot, and succeed in a rapidly changing technology landscape.

To ensure you don't miss other episodes like this, subscribe to our podcast!

Brought to you by TiE Seattle
Hosts: Shirish Nadkarni and Gowri Shankar
Producers: Minee Verma and Eesha Jain
YouTube Channel: https://www.youtube.com/@fromstartuptoexitpodcast

Transcript

SPEAKER_00 0:00

As we dive into the entrepreneurial journey of Praveen Chachadri, the visionary behind Apshi and Thunk AI, discover how we transition from a long tenure on Microsoft to founding successful startups, navigating through multiple pivots, and ultimately being acquired by Google. This episode is parked with insights on innovation, resilience, and the transformative power of AI and business. Welcome to the Startup X podcast. We encourage you to become a Thai member so you can gain access to this great program. To become a member, please visit www.seattle.ti.org.

SPEAKER_05 1:11

After the success of our Microsoft at 50 series, which is the first person account of many senior executives in the early years of the company, we're starting our next series, Startup Spotlight. We're going to feature founders and hear their stories, their journeys, and what they are solving in the new world order of AI. I hope you all enjoy it. We have now actually crossed over 25 episodes, and we really thank you for supporting us through our own journey. It's now available everywhere podcasts can be heard. I hope you all subscribe, tell your friends to subscribe, and enjoy our next series from Startup to Spotlight. Thank you. Now, a word from our sponsor. At JP Morgan Private Bank, wealth is understood to be more than just numbers. It's about creating a legacy, achieving dreams, and securing a family's future. With over$3.1 trillion in client assets under management globally, JP Morgan Private Bank is committed to providing customized financial advice that aligns with each client's unique goals. Clients benefit from a personalized approach, working closely with experts in philanthropy, family office management, fiduciary services, and special advisory services. The firm emphasizes building lasting relationships and ensuring that every financial strategy evolves with the needs of clients and their families. JP Morgan Private Bank not only manages wealth, but also helps clients create a meaningful impact. Whether individuals seek to support charitable causes, manage a family legacy, explore new investment opportunities, the private bank team is there to provide guidance every step of the way. With the presence in key financial markets worldwide, JP Morgan Private Bank offers unparalleled access to global insights and opportunities. Its commitment to excellence and innovation ensures that clients receive the highest level of service and expertise. Time is recognized as precious, and JP Morgan Private Bank values every moment spent with its clients. The mission is to make the financial journey as seamless and rewarding as possible from the initial consultation to the ongoing management of assets. To learn how JPMorgan Private Bank can help clients achieve their goals, visit privatebank.jpmorgan Connect with a dedicated team of specialist. Again, privatebank.jpmorgan.com. Thank you.

SPEAKER_04 4:08

Welcome everyone. Very pleased to uh welcome Praveen Shishadri. He's our first installment in our startup spotlight series. Praveen is a longtime Microsoft, then did a startup called AppSheet, which was acquired by Google, and now has a new startup called Thunk.ai. So welcome Praveen.

SPEAKER_01 4:30

Hey, thanks for having me on the show. Great.

SPEAKER_04 4:32

So there's a lot of interesting stuff, you know, with the work that you did with AppSheet. So we focus uh mainly on that. So first, um, you know, you were at Microsoft for 12 plus years. Yeah. Um, and you know, when you started AppSheet, um, you know, you were middle-aged and uh with a mortgage and a wife and two kids and all that. And what what made you uh decide to do suddenly a startup?

SPEAKER_01 4:56

I wouldn't say it was sudden. I had turned 40. And I think you've all been at my uh some uh you've been in Microsoft, others have too. And it was a contentious place. And um I realized that I was not a good fit for that corporate life. And so of course it paid very well. But uh there was things I was interested in doing, things to explore, ideas to pursue, that were not sure bets, and that's very difficult to do in a corporate environment, I think. You have to have a lot of people aligned behind your ideas and your convictions. It's tough to take a uh you know, fringe bet. And so um I realized if I had, you know, stayed in Microsoft, I'd get to 55, have a lot of money in the bank, and be sort of bitter and have experience, you know, with having sharp elbows. Um pursuing a startup might lead to a different kind of fulfilling experience and it sort of turned out that way. All right, great.

SPEAKER_04 5:53

So you finally made the decision to take the uh the plunge, and as I understand it, um you didn't particularly have a specific idea in mind when you left Microsoft and then um you uh took s uh a couple of pivots to get to finally get to APSC. So tell us about that journey. I think it'll be very interesting for other startup founders to understand, you know, how do you kind of get to the final startup idea?

SPEAKER_01 6:19

I think with the generosity of hindsight, you might say a couple of pivots. In practice, it was at least five pivots. And we didn't have the faintest idea, to be honest, what we were doing in terms of running a business. It was two of us who were both really we were engineers, we knew each other well, but we had never run a business, we'd never really thought of what it takes to run a business. So we didn't have any of that sense. We thought we were good product engineers, which also actually is a questionable um premise, as it turns out. And uh we learned, we had certain ideas about what we wanted to do at the time. I remember leaving Microsoft and thinking, I want to let, I want to allow, empower non-like regular people, non-programmers, to be able to program the internet. That was my broad statement is there's all this stuff on the internet. We have got these powerful computers and mobile devices. It should be possible for anybody to describe some way to program these computers to utilize that and organize it and do stuff with it. And honestly, that was such a vague idea that the fact that we ended up with some kind of platform that vaguely resembles that vague idea through many pivots is shocking. But yeah, that was the vague idea, and we didn't know quite how to what to do with it, how to harness it. And it was sort of super broad and ambitious.

SPEAKER_04 7:34

Got it.

SPEAKER_01 7:34

So how long did that the series of how long did it take for for you to finally come up with the idea for appear we took two years of wandering the wilderness through five pivots, working with it, trying to find maybe we tried to do uh stuff in the consumer space and shopping. We tried to do something with retail. We for about nine months were building apps for trade shows. We try so we just went through so many pivots and we kept rediscovering that eventually we're trying to build applications for some kind of business purpose. And uh, since we were clearly failing at the any business side of this, on the technology side, it seemed like it's not start from scratch every time. There's some platform perhaps here. Let's just fall back to our comfort zone. We are like platform engineers. Let's fall back to something we kind of understand. And so we ended up, which I think I started, I left Microsoft in 2020, early 2012. It was only early 2014 that um we got to this idea of doing AppSheet. And then it took another only October 2014 that we had the first uh beta release of it.

SPEAKER_04 8:34

So what was tell tell us more about AppSheet? What is the basic idea around AppSheet?

SPEAKER_01 8:40

Super simple idea. Um you can build apps from your spreadsheets. And why spreadsheets? Just because people often had some data that they were structured and put in a spreadsheet, it seemed what we discovered from those five pivots, it sounds all like a complete waste of time, but no, what we discovered is there's all of these applications that just look like effectively renderings of structured data and the ability to capture back some structured data through some semantic application. Whether it's a CRM, whether it's some inspection application, whatever it is, they all look like relatively simple database applications. It was super complicated to build mobile apps on Android and iOS with all their weird oddities, offline mode, and so on. So we said, let's platformize all that. Regular people should be able to build these applications. And that was the core concept, and that was still the it's still the core concept, it's still a product part of Google Workspace. The nice thing about it is it was sort of a democratizing concept. There's a lot of people who are not programmers who are able to use that.

SPEAKER_04 9:33

Right. So how so so let's say I have a spreadsheet with some customer data in it, um, and I want my salespeople to update it using the mobile phone. How do I quote unquote program it?

SPEAKER_01 9:47

So this is where the weird thing is we discovered a number of things along the way that now have terms that back then did not have terms. We built a no-code application platform. I remember telling people this, you don't need to be a programmer, and they looked at me like, you must be crazy. What are you talking about? People have tried this sort of thing and it's nonsense, it's never gonna work, etc. But it took till about 2019 for no code to become a buzzword, and everybody wanted to slap that on their existing products. But back in 2015, that's we were building a no-code application platform. The second thing we did was so that's a answer is you just walk up to uh app sheet and you say, I want to build an app mobile application. Here's my spreadsheet, click a button, and it would produce the first cut of the application that worked already. You can install it on your phone, and then you could tweak it and change it and add to it and so on. But it had this magic of in the first two minutes, you clicked, you appointed a spreadsheet, it deduced the structure, figured out the app, it gave you something you can input, you can output, etc. So the other thing that was interesting is we chose to make this a premium product. So just self-service, you walk up to you just sign in, you're free, Gmail, whatever, and you can use the product for free. Some fraction of those users, if they got above a certain kind of usage, we'd say, hey, you need to get to to pay with a credit card. This is now called product-led growth. But back then we didn't.

SPEAKER_03 11:03

Yeah, yeah.

SPEAKER_01 11:04

Right, right. So uh we discovered these things purely by accident.

SPEAKER_04 11:09

Got it. You know, so how long uh so it took you about nine months to get the product into the market.

SPEAKER_01 11:14

We've been building stuff for two years and pivoting, and every time we didn't start from scratch, so there was some basis of code and platform to build applications. Sure. And then to do this new code model took us nine months. We launched, and this is something that most entrepreneurs don't want to hear, I suppose. But your launch of your product is a tree falling in a forest most of the time. And yeah, other than your friends and family, nobody cares. So we launched as a tree falling in a forest. And um, a week later we still had no users. And there was a really weird accident thing happened where we built an add-on into Google Sheets. The Google Sheets has an ecosystem and it had add-ons. And back then, it was quite simple to add one of those things in there, and there were relatively few of it. So we said, Oh, we'll do this add-on to Google Sheets so that you can start there, click a button, and build an app. Took me a day to build. There's only two of us, so I happened to build that thing. Took me a day to build, they approved it within a few days, now it takes many months. And we started to see 10 to 20 people try it every day. That stream became hundreds of people a day. And it's hundreds of people who would try a product. And there were people who actually had some data and were interested enough to say, mobile app for my sheet, let me try it. So that's what got us actually to succeed is because there were these amazing users who would treat a completely broken product, just really broken in various ways. It had some minimal functionality. And instead of walk completely walking away, some some of them would say, Can you do this for me? Can you make it work? Because it's broken right now, but I want it to work. And that's what happened is we had a user community of people who were on our case, and we just decided to love them and react to them and build stuff for them and stand in that loop for five years.

SPEAKER_04 12:52

Got it. So it was really a product-led growth kind of model. To some extent, I mean you uh had a freemium model, offered it for free. Correct. What percentage of people uh converting to uh premium users? Uh three, four percent. Oh, that's a pretty good number. Uh what I normally tell people is if you get between three to five percent converting, that's a pretty good yeah, these were double.

SPEAKER_01 13:15

These were good. These were people who came from a work uh environment. They were not people who were looking at it, they were people who were using spreadsheets at work. And so I think that was very important. But it's more than I would just say product-led growth. The presumption often product-led growth is that we're product geniuses, whoever isn't startup building stuff. And then you're it's a means to acquire customers in product-led growth. Here we were actually uh it was a customer-led product. That's the truth of it, is all the features in the product came from people asking for him. We did not have the wisdom, intelligence, knowledge, nothing to know anything about this stuff. Customers came from every industry and actually 120 countries. 120. Um, so obviously we didn't know nothing about any of this, the any of their domains. What do I know about running a restaurant in Baghdad? But we had a customer running a restaurant in Baghdad who in the middle of summer wants to build an app. We had customers in Africa, we had customers in Canada and everywhere else in different um professions asking for things. So it was quite unique.

SPEAKER_04 14:12

So was your sales model uh primarily B2C, or did you uh hire salespeople calling to enterprises? Because you know, this would be a great, you know, a lot of applications in the enterprise could be built very easily using AppSheet without you know involving IT.

SPEAKER_01 14:26

I would say that uh we did not crack that enough. As in Did you try it, though? We tried. I mean, for a while we were purely self-service. You know, I was kind of an idealistic kind of uh engineer. I said, What do I need salespeople? So there we had a I had I went through that phase. We said those people are adopting this product, we're willing to put in a credit card, they're already using the product, and we'll upsell them. So first salespeople we had was really their only focus is to upsell the people who are already self-service but paying customers. And that actually was a reasonable model. It grew our revenue, uh I think it grew our revenue to about three million ARR. Okay, well so I think about half of that was just purely people putting on credit cards, and another half of it was people we then upsold to some little more of an enterprise plan. Still super low priced compared to you know the big players, but but we then I think eventually, I think 2019, we erased a series A. At that point, we said, all right, let's go and get up more of a professional sales team with you know sales leader and SDRs and all of this stuff. Uh again, though, most of our sales were the people who are already using the product. So we had that good fortune of that, and we were able to sort of double revenue that year, but then we got acquired, and so then it's a whole other story once you're part of the a big the big uh big umbrella like Google Cloud.

SPEAKER_04 15:46

Right. So how uh how many users did you ultimately have?

SPEAKER_01 15:50

I think uh at the time we were acquired, we had uh 8,000 paying accounts. 8,000 paying accounts, okay. But most of them small. I mean I was I one of the principles I had rightly or wrongly is more or less I treated the person paying five bucks and the person paying uh 50,000 the same. You're paying customers, they deserve, you know, uh responsible, professional, you know, accountable treatment from us. And uh so while I say yeah, there's eight thousand distribution, there were certainly some at the high end of that, some enterprise customers. The the customer who paid ten dollars were still treated with you know as an important customer.

SPEAKER_04 16:25

So let's talk about your funding journey. You went through, you know, you raised some initial funding for your company, uh, friends and family, etc. And then C-Down and CDZ. So take us through that. You've written about it. Uh yeah, I've written about it. What are some of the lessons you learned from that?

SPEAKER_01 16:40

I would say if anybody listens to this, I suck at raising money. So I'm not a good person to take advice from. Except you can take solace from the fact that you can suck at raising money and still survive. That's the that would say the high-level uh takeaway. My general principles are forget about everybody else, because you hear about everybody else raising large sums of money, and you think, oh my God, well, everybody's raising this much and that much of this valuation. It doesn't matter. You got a business to run. And so the question is, what money do you need for your business to succeed? My principle was always, perhaps by circumstance, what's the least money you need to succeed? Perhaps if somebody had written me, like, I don't know, soft bank and decided to write me some huge check at the start, I'd be a different person. But nobody wrote us a check at the start, nobody's prepared to. Um and so we learned to just basically be super frugal. I remember eventually you got acquired one of our VCs. We had eventually we raised a CVC, and one of our VCs, Greg Papadopoulos, who's at NEA, told me, Greg's a great guy, said, Raveen, you should be really proud. You've been so frugal and spending money, and you've just so you've done this on with you called it sipping capital. That is, you just had enough to sip at it.

SPEAKER_04 17:50

Couldn't you couldn't take a swig of the capital, you know, it's sip at But could you uh could you have uh I wonder though, uh if that was the right strategy. Maybe um if you had uh invested in a business partner sales strategy, enterprise sales strategy, I would have been much bigger. Ultimately you may have been still bought by Google, but it would have been a different price.

SPEAKER_01 18:11

You're totally right. I don't claim this actually, this is why I started out by saying I started raising money and and so on. I do feel that business were undercapitalized to invested in and therefore it limited its growth. If you have the ability, at the points at which the product is ready and can grow, it needs to have that fuel to feed it. Well, at the points where the product is not ready, you only need to feel the fuel to build the product. And judging that is super difficult. I'm not a fan of having large teams or pursuit until you need a small core team to build the core product. It's not expensive. It's expensive for those folks with but they take equity. And so what you really you do not want to sell equity to investors at low valuations when you're building the initial product. It's a very poor use of money. But you do want to raise money, adequate money for growth. And exactly what that ratio is tough to tell. We managed eventually to raise money for growth. It helped us grow, but probably was too late in the cycle. It pushes in hindsight, it would be better to have raised it earlier.

SPEAKER_04 19:14

How much series A did it raise?

SPEAKER_01 19:16

We raised a 15 million series. 15 million. 15.15.

SPEAKER_04 19:20

That's a substantial amount of money.

SPEAKER_01 19:22

Yeah. Before that, we had raised about three and a half million on safe notes. And with that, we had run six years. That's about generating three million in ARR at the end of that period. So we basically were break-even with about a 15-person team. By the time we raised a CDZ, once we raise that money, then we were like, okay, hey, then we the team grew to 45 people, which we doubled our revenue in in a year. Um, but um, and then we got acquired. So it puts you on a different trajectory. One point to close on the thing, super, super important, not what money you got, but who you get it from. And so I'd say that's a whole other dimension to the people investing in your company are it's kind of like being in a in a long-term relationship. And so it has to be people aligned on values and aligned on your goals and uh incented to do who believe in doing the right thing as defined in various tricky situations. Um very, very important to have the right people. Sorry, Gaurie, go ahead.

SPEAKER_05 20:32

No, no, no, Prami, that's actually quite good because when Shirish and I talked to founders today, we advise them that early capital is good, but it's also you gotta think in seven year increments, not two, three. I mean, I said that's uh that's a seven to ten years is a long relationship in terms of a professional one, right? Yeah. And this is one of those you take home and bring back. It's not like you can switch off at five o'clock and go back, right?

SPEAKER_01 21:01

Correct. There's also those subtleties which I was not at all familiar with, but I have this now a principle. You take a big check from a big fund and you take a small check from a small fund. You never take a small check from a big fund.

SPEAKER_05 21:12

Even if offered to you.

SPEAKER_01 21:15

And this is not at all obvious to a first-time entrepreneur, first-time entrepreneurs like, oh, let's pick up a favorite big firm. I don't know, score, Axel, somebody says, I'll give you a small check. You may feel, wow, this famous fund wants to give me a small check. Super bad idea for various reasons. You know, that's I think important principle. If you're collecting a small amount, raising a small amount of money, go to angel investors and the smallest funds.

SPEAKER_05 21:39

Um yeah, that's very good advice. That's very good. So let's uh let's switch to your current startup, right? So now you have a track record of starting one, pivoting into a success and getting acquired by Google. So this go around after your stint at Google, did you land on the idea directly, what is it that you're doing? And did you land it right away?

SPEAKER_01 22:04

Well, what we're doing first, the company is called tunk.ai. It's like think, but it's taught already. So it's tunk.ai. And we're focused on utilizing the modern Gen AI to improve business productivity. Basically we learned a bunch about business productivity by doing AppSheet. Along comes this huge transformative wave of AI. And the question obviously it has various consumer applications. And the question we asked was can it be used to automate work? So a little more than two years ago, I quit Google one of my co-founders Tony who had been at AppSheet also quit Google. Our third co-founder quit her job at uh Facebook. And um we said well let's try to apply the Gen AI technology to help people automate work and drive business productivity. Now fashionably this would be called AI agent automation. When you started didn't have the term like what are you trying to do? Anyway, so we've also trying to figure out what are we trying to do. And so it's been a sort of a very interesting two and a half years. I'd say the core idea has been consistent. The technology has changed a lot in the last two and a half years. And so it is a space where there's a lot of anticipation and judgment that's needed far more than anything else that I've kind of lived through and seen because there's a lot of unknowns and there's a there's a lot of their pace of technical change is very fast. But that's a broad space we're in um and we've been doing it for a little more than two years. When we've built a platform it's sort of in some sense in spirit it's similar to AppSheet in spirit and it's a no code platform to describe workful applications that you want AI to run. And AI runs them and runs them reliably.

SPEAKER_05 23:42

So are you selling this time to enterprises or to uh uh smaller businesses?

SPEAKER_01 23:48

One of the um some of the hubris that comes you mentioned hey you had a success before and uh all these startup journeys are these fluctuations between uh hubris reality maybe some hubris again okay so I would treat my app feed journey as left Microsoft hubris think you're great turns out you suck at everything and then you sort of dig your way out of that say it's start doing something useful for people and so on. So likewise I think we sort of uh left Google to start this thing and just startup you have to start with some hubris. And so the hubris here is hey we want to do something transformative. What we really want to do is actually AI should make it possible to truly deliver a billion people being able to build things that we can productively run. And they don't need to be able to specify that some incredible granular detail but AI should fill in all the gaps into that work. So that's our starting hubris filled thesis if you wish and so our eyes are wide you know our appetite is ha ha ha so that's sort of where we're uh coming from so to your question enterprise or not the fact of the matter is in the moment there's tremendous amount of noise around AI and AI agents. There's a lot of confusion around it there's a massive opportunity for S and B and there's a massive opportunity for enterprise. And the question really is are we building a platform that's only for one or the other or both we committed this time we're going to build a platform for both. We've built this platform so it can be installed on premise a bunch of containers in an isolated instance of an enterprise wanted. But there's also a multi-tenant thing in the cloud it can be self-service. So we've intentionally built for both now who's going to adopt it first is an open question. A year ago it sure seemed like the enterprise was going to adopt it first because in all the enterprise companies from the board down there's a mandate thou shalt use AI. It's clearly it's this transformative wave what are we doing with AI is a question driven from the board down to the CIO. Whereas in the small and medium business any of these things is bottom up it's not there's no board driving this. It's really about is a practical value that can give us an edge can be and so on. And so when there's confusion about what AI can do and a lot of hype and so on, SMB is typically not at the forefront of that. So I would say in the last 12 months the enterprise customers have been the ones who have already got the mandate to say we should be looking at AI agents. So we've focused a little bit more on the enterprise customer in the last 12 months just to work with design partners, work closely with a few to hone the platform. But it's an interesting moment in time because a lot of the hype is now being also tempered with oh is it overhyped because there's now this large number of voices calling out some obvious known technical things which is but is this AI reliable? How can we actually put any of this stuff into production when it's all mixed with the demos. So a lot of the enterprise companies seem to have slowed down and in terms of they're running to the reality reality checks. Whereas I think SMBs had 12 months to sort of get used to the idea that you can automate things with AI. And so there's opportunity on both sides. And we'll go wherever there's a pull really be honest.

SPEAKER_05 26:55

Right, right. So one of the theses I have and I want to sort of dig into how your company is going to solve it SMBs really did not have as much productivity tools like enterprises did. Right? Enterprise along the way I mean your previous company was was was a productivity tool that was available first time for the SMB but there was lots of other productivity tools available for the enterprise all along there was a whole bunch of consultants who are automating it was coming along right now AI it seems like it's democratizing tools for all I mean the fact that I could imagine writing poetry it's in itself is the first sign that AI is democratizing that. So is the opportunity that what you're building the AI agent tool set for SMB a real one that they could adopt faster than in App Sheet they were discovering hey I I don't know how to build a build a you know an app for a mobile device and this allows me to do it. Now are you bringing it to that point where like hey I could do it because enterprises have all the help they have they have Google, Microsoft, Salesforce, everybody calling on them to say I'll help you with AI agents. I don't think they're calling SMB This is true.

SPEAKER_01 28:13

To sort of unpack your question on this aspects of it yes we're building the the product we have is something that an SMB customer with relatively little help can pick up and create AI automations with. So that's a and that's definitely the intention. There's a separate question of whether or not enterprise customers are well served by the product offerings of the legacy vendors. To me it feels like having worked at Microsoft, worked at Google, they're all great companies. But the real question is is their suite of products really what you would build in an AI first world? I mean that we all are forced to use a set of products around office productivity that look effectively like Microsoft Office from the 90s. They just happen to be loaded through a browser but I got all my email in one place I got my file system in one place I got my slides here I got my sheets there. That's a sort of a strange way to organize your information if AI was going to look at it. It just happened to be the way stuff was organized in the early 90s. But you see the legacy vendors have basically locked, have made their businesses around building siloed product and whether it's Salesforce or whether it's Microsoft or whoever else who is now saying hey I'll give you AI there is this question of If we fundamentally set up all these enterprise customers to have a broken and inefficient system that no amount of icing on that broken system, AI icing is not going to solve for so at some point I do strongly believe at some point this AI wave is going to break these product silos because the AI logic has to work across the product silos to do intelligent things in some particular domain. If I'm working with customers I got some information messy and some are sitting in my email and some are sitting in my file system and I got to tie those things together to do anything meaningful. We as humans tie it together in our brains none of the vendors traditionally have built their systems this way. So there's an opportunity but it is a long shot. So we have to acknowledge that and it comes back to when you do a startup do you pursue the the long shot opportunities also? And those are the ones that get you big wins.

SPEAKER_05 30:22

Right, right. So you mentioned when you start AppSheet hey you are great engineers but really didn't know how to run a business. You ran one you successfully exited it worked in a big corporation that might have shown you how to run or how not to run but whatever you learned out of it. You came out this go around if if I'm an entrepreneur listening to you, if you could address three or four areas right how do you know you've achieved say product market fit you have a go-to market plan. Your pricing is reflective of the value you're giving and all the other business stuff that still needs to be done. I mean it's not going to go away you still have to kind of do the business stuff of all this so in round two now that you're a serial entrepreneur how are you how would you advise on these uh threads to a to a first time take back to Praveen of 2012 who said I'm going to start a company what would you be telling Praveen of 2012?

SPEAKER_01 31:22

Let me say this uh very few people can fundamentally change and create generate new strengths they didn't have when they were you know by the time you're 40 or whatever and I'm 55. So you're unlikely I'm unlikely to have become a genius at business and I'm not so it's useful to have a group of uh co-founders or executives who balance out the needs of the company. I would say that's a really important thing. And each person if you're a founder focus on your strengths. I think that's my main um thought in this is that's one meta thought. I'm fundamentally product and engineer and I like to look at the world through that lens. And it's therefore useful to pursue a business which aligns at least with the majority of the founders in terms of how they view the world. Because we view the world the product first lens we believe that if we build a good product yes we know there are successful businesses built with mediocre products but we have no shot at running such a business. Our only shot is if we build a good product and maybe we can run it as a mediocre business people but the product has to be great. And so I think that's our first observation of about ourselves we really believe that the product must be compelling and that's a B, I developed a very very very strong um uh value system around prioritizing customers. And I think that's a long-term value uh you know principle rather than a doesn't translate necessarily short-term revenue and so on. But I think so at least the way we're thinking about this is build the best product you can build. Keep refining it. Have a maniacal focus on your customers and making sure you value them, respect them, make them happy. And the business part of this will be reasonably straightforward if you built a compelling product for a real need. If it turns out you have to do special business things, have true business unique insights and so on, I certainly don't have them. And so that can't be part of my strategy. Hopefully we will have other people in the team who will be able to squeeze the most out of their business choices and business strategy and so on. But I don't plan on it. I plan on a really great product um do put in front of customers and deliver most of the value to them. In other words I tell our customers you mentioned pricing but right now AI pricing big open topic so customers who work with us will do a free proof of concept they will say okay it's time to get into a contract they say tell me about your pricing and I say a very simple thing here's the principle behind up pricing. You should be generating value there should be value being generated by you doing AI automation on this most of that value should accrue to you fraction of that should accrue to us. 80% should accrue to you 20% should accrue to us that's the principle through which the lens through which we approach that right we have to then translate that into something practical. So as a practical thing we might say initially I know you're only doing these three projects you're not doing more whatever so here's what we propose as a platform price. Is that good for your one? We will figure this out together now you may look at this and say this is insane. This doesn't sound like a business person at all but you notice it's very much aligned with a person who basically says I'm valuing my customer and I want that just to be a customer driven product. And so that's the sort of approach we have on this at the moment and maybe it's uninformed or maybe it is just idealistic but that's the way we're proceeding.

SPEAKER_05 34:37

So in your current uh that that is actually good advice by the way because the discovery of what you're talking about with your customer I think is extremely important when the underlying technology is shifting at a pace that is not seen by most of humanity in general. Because you know all in many cases the underlying technology was solid then you might have innovated like Uber was an innovation on location but location was well known declaration of location was well known way before Uber came along but this is very different right so I like that advice you gave to our audience.

SPEAKER_01 35:15

Yeah I also say I can say this as a small business sort of I would say startup kind of person you never want to be in a situation where your customers feel that you're overpriced. In fact I feel the software industry in general is making a big mistake. Say this having talked to enough especially how enterprise customers are super fleeced to be honest. They're charge products at ridiculous prices and sometimes get locked into them it's countless number of customers enterprise customers who get into there a software vendor sues them it's sort of ridiculous forces them to do uh upgrades and charges them back fees and people there's sort of this almost this mentality of if you get into the enterprise market then you get to generate lots of revenue because you can overprice these customers so I actually are would much rather be in the situation where the customer basically says you're giving me a very fair price principle always used to be that any customer ever basically said I'm happy the product we basically say we'll refund your entire money entire money right there almost no service that does this right you sign up for a service and I'm not happy with it I paid my month I paid my year it's never coming back. Yeah but that's such a bad approach for a startup to have startup really has got to think about the next decade of where these customers respect you even if they walked away did they respect you.

SPEAKER_05 36:35

Right, right right um so based on your previous experience was your go-to market different this time around or similar I mean you had this sort of attachment into Google Sheets that yes propelled it. Was that something you could do here?

SPEAKER_01 36:50

We could try we've tried I don't think that's going to work in this case at the moment. The point is there's just a lot of factors that have to align when we did App Sheet the iPhone had been out for six years at that point in time. Everybody knew what a mobile app was everybody could imagine what features they wanted. They just couldn't build it it was a relatively simple problem. So there's no problem describing the value prop of having the out the result. The biggest problem everybody's having around AI today is beyond ChatGPT nobody's able to basically say damn it I want this AI thing. They could say I could speculate that maybe this would be cool. Everybody has this problem whether it's Microsoft, Google, whoever is selling AI products right now has this problem of people like okay it looks nice but do I really want it? Is it really fundamentally going to change the game? Nobody so there's that unknown yet at the same time you've got this ridiculous amount of oversaturation of the market in terms of the hype the marketing we don't have to educate anybody on why AI in terms of what is AI. In fact everybody seems has a different notion of what is AI partially right and partially whatever. But so tremendous amount of pre-marketing but not a whole lot of value proposition that's sunk into the customer. So it's not as simple as put your product out into the stream and the fish will bite. One we have to go through this discovery process of working closely with a small number of customers to build a product that really matters to them. And that's kind of approach for us is we work with our 10 customers we work super closely with them we meet with them every week engage with them super closely and make sure the product works for them. And that turns out to be super non-trivial just because of where AI technology is today. And the non-trivial part by the way it's not that you can't invoke in language model and get results and so on. But making it work reliably is a difficult thing. And most of the in fact all the AI agent products from vendors um there's various benchmarks and things they say like 30% reliable 60% reliable we have vendors telling customers just get used to it you're not going to get more than 70% reliable but just find find applications where that's okay. And then you ask yourself why would why wouldn't the customers be a little reticent to go over that. So we're sort of focused on how do we make this an automated process 98% plus reliable which should be at least as good as or better than human reliability on some of these tedious processes. That turns out to be super technically deep problem so it feeds on desirable technically deep product and so that's where we're at.

SPEAKER_05 39:22

Got it. So let me sort of go to go down this path a little bit with you on you said have good co-founders because you have your strengths, right? The others could bring other strengths are complimenting. Where do you stand on mentors and advisors and who would you pick if the world was the oyster who would you pick and say I want this person in my corner? Not because they want to be or you want that just as a metaphorically speaking.

SPEAKER_01 39:45

So there's two separate things I think one is who would the would benefit the company by them opening doors for the company getting introductions and so on. That's less about personal mentorship. It's just more about you know you know I'd say somebody who's you know run a big successful enterprise before if they were sitting on our board and felt incentive to basically give us introductions and tell other people how great we were of course it would benefit us. So that's I would just say that's a feels like a very transactional kind of response to your question. I think your question has a bit more about who can provide mentorship and guidance to founders as people and go through the founding journey and so on. And there it's actually weirdly people who can who AU must respect already and they must know you already because you don't have the bandwidth in this process to have to get to know the person very deeply. So if it's a good example somebody who you worked with closely before maybe you was a manager mentor something before so they say hey I already know you Braveen I know your strengths and weaknesses and so I can advise but and I'm fond of you so I can advise you in a way that you don't take offense when I have to push on something. You're far you you can have people who you seek that counsel and there's a lot of people giving you free advice in this world and generously that's a beautiful thing about startup ecosystem people are super generous well intentionally generous so you don't need to have them as a mentor but almost nobody's incentive to tell you Praveen you're you're screwing up nobody's incentive to do this. So if there's somebody who you trust enough and they have a strong enough relationship with you they know it's not going to be broken by them saying this to you that is valuable. So on AppSheet we had the you know at one point when we raised my VC money we had a board so we had two of us as founders on the board we had two VCs and there was a neutral uh board member and the neutral board member was uh Pino Spiro who was he had a long time ago been my um manager at Microsoft and he was a technical fellow retired whatever so he'd been retired for a long time and so he's he's he's my friend but uh Spiro was able to tell me things like you know Brave you falling into that trap you've I've seen you fall into before where you get you know super focused on this one thing and you lose sight of this or you sure about this so and I was prepared to hear it from him also. And so that sort of mentorship I think is very valuable in whichever dimension it could be purely about just you know handling stresses and things or it could be about something more specific like business or technical or whatever. And it's tough to find there's relatively few for any of us there's relatively few people in the world we have those relationships with and if you're fortunate enough that they're willing to spend time on engaging with your startup to mentor you on this then you're very fortunate.

SPEAKER_05 42:36

So just as a blast uh as we as we wrap this up would you see this go around compared to the first one? I mean and there was seemingly a a 10 year gap 2012 and around 2022 when you decided to jump off so there's certainly an age gap and difference and all that but you've now that you've uh you know you've done this the second go around was it different this time or were you very clear and sure because again you had this crossroads you could have stayed in Google big bank account uh like you said and maybe miserable maybe less miserable I don't know but but uh you made a very intentional decision to come back in I'll say this uh uh first to why I'm 55 so we made this decision 53 the decisions you make at 40 you have more runway than decisions make at 55.

SPEAKER_01 43:36

you have less uh there's less at stake than as the previous company uh company you know I had my kids were in middle school and my wife went to medical school I did a startup and neither of us took a salary for five years or so it's a sort of a bigger riskier decision. This one's more a decision what you do with your life because you know uh you count the you're fortunate to be uh healthy and able to do this and the question becomes um if you're going to commit 10 years Your life to work on something, is it useful? And I have to uh, you know, and so I think it was worth it did occur, uh it felt to me that this wave of AI was the single most exciting thing and transformative thing I've ever seen, and has happened in in computation since the first computer. So I was born in 1970, first computers came out before that, and since 1970, there's not been a technical breakthrough of this potential for change and proceeding at this pace. So, what the hell? Why did I go get a computer science degree, PhD, teach, work on all this stuff for 20-something years, work on all the stuff, and then at this moment say, nah, sorry, I'm just gonna spend my time retired. I like working, I like working my 80 hours a week. That's what I do. I write code every night, I write code every day. And that's so I this is what this is my retirement. It's you're swinging for the fences for this, it's great. So to me, it's like uh it'd be crazy not to pursue this moment. So that's sort of simple, it became a very, very simple decision to jump in. Once you jump in, then of course it becomes like, hey, there's a lot, it's not just about you individually, it's about the team collectively, it's about investors, all of those things. But that personal decision, super, super simple and stress-free because what's your your opportunity cost at this age is low, except for clock sticking. Biological clock is ticking, so you know that opportunity cost is super high. But other than that, you know, um, it's not a yeah. So I would say emotionally it's much easier.

SPEAKER_05 45:29

So do you have investors now on this one already? Yeah, or is it all self-funded?

SPEAKER_01 45:34

We self-funded initially, and then we raised a safe note. So we just made sure, like I say, we made sure I just follow the same principles I wrote, which is raised smallish amount of money from small investors and uh enough to make sure we can build a product. And we're pretty much there, we still have quite a bit of money in the bank. So we should be able to get to demonstrating product market fit this year and then go raise money to go uh scale it.

SPEAKER_05 46:00

But since you have already executed an exit, right, from AppSheet, right? The path to for investors to trust you is different now. Because you know, you came from Microsoft to a to a AppSheet, but now you've exited. Yeah, you had you had a pit stop in Google, but everybody knows you can sell into Google, right? So has that changed investors' interaction with you? Or people saying, hey, I want to in on it. I don't know, I don't care what you're building.

SPEAKER_01 46:30

Nobody said that to me, so not quite there. But your question is has it uh changed the investor approach? Yes. As in the way I describe it to some people is earlier it was super difficult to even get the conversation. No, it's not difficult to get the conversation. It doesn't mean that it's easy to get the check, but um so progress. It's just that what I went back to our ambitions are bigger. So I think that there's that also. So uh there isn't a room for here to do a slow burn for say six, seven years. There's not that that's not realistic. So you have to work faster, raise more money earlier. So that that trajectory is different. And also doesn't change the fact that at the end of the day I suck at raising money. So there's that too. That that has not changed. Um yeah. The big difference is the first time around I spent a lot of time chasing money, thinking that was a demonstration of success of progress. Here I spent no time chasing money. None.

SPEAKER_03 47:23

That's a good thing.

SPEAKER_01 47:24

When it looks like we should be able to raise money, we'll try to raise money. But right now, we we have to focus on the product and the customer. And until then, zero distraction, talking to VCs, talking to whatever. It's just not not an issue. And that's the luxury we have right now, and so it's super wonderful. It's not a confidence you could have had unless I'd gone through that once before.

unknown 47:43

Right.

SPEAKER_05 47:44

Pravae, thanks. That was a like a master class on do's and don'ts of uh of a startup, right? You can be an engineer and have a successful exit. That's that's your first thing. You keep saying you suck at it. I don't know, you exit it.

SPEAKER_01 48:00

There's there's many don'ts, by the way, so but uh we didn't cover all that the that would be a whole one hour of all the screw ups uh that we did and you shouldn't do. But yeah, but very few of them are terminal. You know, just keep at it. That's sort of maybe uh mentality is just one foot in front of the other and keep at it.

SPEAKER_05 48:20

Sometimes it sometimes we'll bring you back for uh for the part two. So back to you.

SPEAKER_03 48:28

All right.

SPEAKER_04 48:29

Thank you so much, Praveen. Uh great conversation. And I will also link to your uh medium post, quite uh uh illuminatory. Uh thanks for sharing all your wisdom with the startup community and wish you all the best with thug.ai.

SPEAKER_01 48:45

Absolutely. Thank you both for the opportunity to talk. It's great, and I look forward to seeing the rest of the series also. Okay, great.

SPEAKER_05 48:53

Thank you. Thank you for listening to our podcast from Startup Exit brought to you by DAI Seattle. Assisting in production today are Isha Jain and Mini Varma. Please subscribe to our podcast and rate our podcast wherever you listen to them. Hope you enjoyed it.