“If there’s no friction, there’s no market.”

This week, host Conor Bronsdon chats with Dheeraj Pandey, CEO and co-founder of DevRev. Dheeraj shares his journey from leading Nutanix to a valuation of $16 billion, to now striving to build another unicorn with DevRev.

Dheeraj opens up about the lessons learned from his entrepreneurship, emphasizing the significance of having a holistic approach to building products and companies. He explores how his experiences at Nutanix have shaped his vision at DevRev, especially in terms of integrating AI as a core component rather than a mere add-on.

The discussion also covers a range of topics, including the power of AI in business operations, customer-centric design, and the importance of balancing innovative technology with intuitive user experience.

Topics:

  • 02:49 What inspired you to start DevRev?
  • 08:38 How did you decide not to take a bolt on approach to AI?
  • 11:30 What can other entrepreneurs learn from you about levelling up their abilities?
  • 16:42 How are you approaching scaling up DevRel?
  • 28:11 How is DevRev thinking about AI driven design?
  • 44:52 What's happening with DevRev's leadership in AI?

Links:

Dheeraj: 0:00

And the, way we dealt with not just making the product very easy to use, but also making customer centric customer service so delightful. We said, we got to really make this into an operating system. And that's how came DevRev, obviously timing is everything. You know, we figured the killer app for AI is customer support. And yet we couldn't be a bolt on on top of existing systems. We had to really go and do this, uh, fundamentally differently with, uh, cloud native and AI native principles

0:30

Developer productivity can make or break whether companies deliver value to their customers. But are you tracking the right metrics that truly make a difference? Understanding the right productivity metrics can be the difference between hitting your goals and falling behind. To highlight how you can adjust your approach to both measure what matters and identify the right corrective strategies. LinearB just released the engineering leader's guide to accelerating developer productivity. Download the guide at the link of the show notes and take the steps you need to improve your organization's developer productivity today

Conor Bronsdon: 1:05

Well, welcome back to Dev Interrupted everyone. I'm your host, Conor Bronsdon. And today I'm thrilled to be joined by Dheeraj Pandey. Dheeraj is the CEO and co founder of DevRev, a unicorn. And he's the former CEO of Nutanix, which he led to a 16 billion valuation and a successful IPO. Dheeraj's entrepreneurial journey goes from Nutanix to DevRev and is marked by innovation and a deep understanding of how AI has transformed businesses already and how it can transform business operations moving forward. Today, we're going to discuss his approach, the lessons he's learned, and how he's applying a design led approach to AI at DevRev. To build that unicorn company and challenge giants like Salesforce and Jira. Excited to talk about his counter positioning and understand more about that. Dheeraj, welcome to the program.

Dheeraj: 1:50

Thank you. Thank you, Conor. A pleasure to be here. Dan

Conor Bronsdon: 1:54

because it's, it's always great to learn from entrepreneurs like yourself who've not only led one successful company, but now are doing it again, uh, because there are clearly repeatable lessons and things you've learned along your journey. But before we dive into some of those lessons and your approach, I do want to remind our listeners. I know they hate to hear it from me every week. But if you're enjoying this episode, if you like guests like Dheeraj, if you want to learn more about these kinds of topics, please take a moment to rate and review our podcast on your podcasting app of choice, whether that's Stitcher, Spotify, Overcast, Apple Podcasts, these ratings and reviews are how people find the show and they matter a lot. It's an immense help. We deeply appreciate hearing from you, our listeners. Uh, with that said, though, we're going to dive into the content. Dheeraj, you built Nutanix into a 16 billion company, and now you've founded DevRev and it's valuation that sounds like is unicorn status. Not to mention, you've also been a board member at Adobe, uh, you're a former VP of engineering and much more. What inspired you to pull away from all these other successes and build DevRev?

Dheeraj: 2:55

Yeah, thank you for the introduction. I mean, that 16 billion has been a labor of love for a lot of people. Timing wise, it just happened to be at the right time in the right place. Obviously, I had two other co founders who've also built great companies over time. and just the folks who've actually been doing it ever since I left about three and a half years ago. So, uh, it's taken a village to build that. You know, when I was doing my first company in 2009, um, the itch was there. I figured that there's a little bit of an intuition, the gut about really going and building products that people like. there's a really good article in The Economist in 2015, uh, when Satya was taking over, Microsoft and, you know, they, talked a lot about humility and fierce resolve that really made Microsoft, Microsoft, uh, you know, maybe 40 years into its journey. And one of the big things they talked about is how they said, look, we're going to have to build things that people like, as opposed to put this strategy tax of Windows and Windows phone. And like, everything has to be built here. Like, no, no, no. We got to really build apps that people might actually like on their operating systems. Microsoft Mechanics www. microsoft. com And to me, that was telling because, uh, it just spoke volumes about, first of all, what it takes to reinvent yourself, but also, how market is king. The market is king, you know, there's nothing more powerful than the market. It's like the ocean and you don't go and surf on the ocean against the tide. You need to actually be with the tide. So in 2009, that journey was about building distributed systems for, uh, for the masses because distributed systems were still confined to Consumer cloud and cloud companies were trying to do this, within their own premises. But he said, if you have to really take this and democratize it, what would it take? And that's what Nutanix is all about. Like if you have to democratize distributed systems, uh, what it meant to have people not buy proprietary, expensive hardware, you know, is there a way to do this? And there was a killer app. The killer app was virtualization. You know, we didn't know much about virtualization and we learned everything about virtualization and what it means to take physical machines into the world of, into the realm of virtual and build a great, delightful experiences on top of Now, the first five years of Nutanix, the product was very difficult, like really, really hard because we were embracing commodity servers, commodity hardware. The things that the public cloud could do. Folks were comfortable with because they didn't have to ship all this stuff to anybody else's premises. We had to ship all this stuff to somebody else's premises. And then we had to support this remotely as opposed to the way the public cloud companies or the consumer cloud companies were doing it with tons of people. I think that was the hard part. Of course, VMware of an operating system underneath us was also really hard because they had never seen as big an app as what we had actually come to build about data management and distributed systems on top of VMware. So the way we had to compensate for that difficulty, the fact that the product was actually fragile was through customer support. We said, look, we've got to really bring in the human element to this technological breakthrough that we're trying to actually achieve in the market and this yin and yang was creating a product with customers, you know, was this yin and yang and customer support was about being customer centric, you know, leading with technology was about being product led and I just, you know, realized that we had to do this differently than what we were used to. Was about ticketing, like okay, customer support is not just about ticketing. It's really about an end to end collaboration between the market, the end users, the buyers, and you know, the support people, but also product managers and developers. So what does it mean to create this real time highway between the end users and the developers and everybody else in the middle. We used a lot of Slack. In 2014, we were one of the earliest users of saying, look, let's do real time collaboration. You know, we did, uh, almost 7 8 billion in those, uh, 8 years of selling between 2012 and 2020. And almost half of it came from a handful of customers, you know. Uh, hundreds of customers, you know, and to really do that, we had to build this, uh, you know, sort of collaboration super highway and way more than what support tools had ever done. So we built this along the way. Our net promoter score was very high, you know, 90 plus for seven years. And so it was a COVID hit. I'm like, okay, uh, this thing is here to stay. This idea of digitization and making things more invisible is here to stay. More and more things on prem will actually end up in the public cloud. What is the one thing that all this will mean? And I'm like, okay, we've learned so much about customer support. in the last, uh, 11, 12 years of when I was building Nutanix, let's make this into an operating system. What does it mean for every company to be as customer centric as we were? And remember we were dealing with, uh, big behemoths, you know, and this was a hundred to 200 billion infrastructure market that had established competitors. And the, way we dealt with not just making the product very easy to use, but also making customer centric customer service so delightful. We said, we got to really make this into an operating system. And that's how came DevRev, obviously timing is everything. You know, we figured the killer app for AI is customer support. And yet we couldn't be a bolt on on top of existing systems. We had to really go and do this, uh, fundamentally differently with, uh, cloud native and AI native principles.

Conor Bronsdon: 8:39

It's really interesting you mentioned that bolt on principle because it seems like a lot of companies today are taking a bolt on approach to AI instead of building from first principles and having embedded within the design from the ground up and you very much have taken the opposite approach it seems.

Dheeraj: 8:54

Yeah, and look, I mean, when we were building our last company, we knew that if there's no friction, There's no market. I mean, you can be building a quick 30, 40, 50 million dollar company, but if you have to go above 100 million, you know, most of these companies, they become one of those companies. Like, now what? Like, where's the time? Uh, they're looking down the barrel and there's nothing to be found, uh, because they didn't go and displace something. I mean, you don't just create new categories and new markets which are accretive in terms of budgets and, oh, well, everything is infinite. No, you've got to displace something and then make it better as well. So this idea of a system of record was very key to us. And, When we were doing this in 2012, 2015, 2016, that whole era, this idea of hyperconvergence, where we're trying to converge a lot of things in pure software. There's lots of me too companies that were just saying, let's not take the big problem head on, which is displacing incumbents. We can't because it's too risky. And by the way, Anyway, the exit that we are really pursuing is kind of being acquired by one of these bigger guys. And we said, no, no, we got to really take this. And only two companies survived. We and Pure Storage were the two companies that survived this because we said we got to go displace incumbents. And this time around, we had to do the same thing. It's not just a system of intelligence that It's a bolt on on top of some system of engagement that's a bolt on on top of a big system of record where the real money sinks in the system of record. It's kind of the hardware of business software, you know. So we had to really take all these three layers, the system record layer, which we have to go displace, enough Zendesk, enough Salesforce Service Cloud, enough Atlassian Service Desk and so on. We had to go and displace these things and then bring engagement and intelligence on top of it. That's the only way to build a billion dollar revenue.

Conor Bronsdon: 10:54

This is great stuff because it's really clear that your experience building a billion dollar revenue company at Nutanix has really shaped your approach to DevRev. Not only Did it inspire a journey there saying, Hey, I think customer support is a huge opportunity. There's more opportunity for end to end experiences. We need to engage everyone with the customer and also like, Hey, let's integrate this incoming technology wave. But you've learned these other key lessons that I'm hearing from you as an entrepreneur, like that responsiveness to the markets, that, deep focus on customers, uh, and on your product. Like as I'm sure we've heard the Jeff Bezos quote everywhere, you know, what are you doing to make your beer taste better? I'm paraphrasing. What are some of the other key lessons you've learned along your entrepreneurial journey that folks who are listening to the audience who are maybe thinking about becoming entrepreneurs or are first time entrepreneurs and are trying to level up, what, what can they learn from you?

Dheeraj: 11:41

Well, I think the big piece of this is people, you know, how do you really, uh, don't shortchange people along the way, because, you know, uh, the people, your employees are also the market and the market knows more than you think they know, you know, uh, this is equally true for product quality. Authentic customer service. All this stuff is about realizing that, you know, you can just go twist market in your direction. You can just go and shortchange them. And people, employees are like that as well. You have to be very authentic with them. You need to know how to let go of things. You know, letting go of things is one of the hardest things that entrepreneurs, um, struggle with. Um, I mean, I struggle with this too, you know, one thing we did well was hire good. Leaders and create new leaders. Uh, also young leaders. They didn't have to be folks who are just with immense amounts of experience because sometimes it's very hard for them to reset themselves, to roll up their sleeves and really, you know, work on a new brand, build a new brand. It takes a lot to build a new brand from scratch. So knowing what it means to balance, uh, the fountain of youth with, uh, the excellence of, experience, you know, all this stuff needs to come together. I think the most important thing though, uh, is how do you weave work and life together? Because many times I see entrepreneurs really swing the pendulum too much to one side, or sometimes it's just a lifestyle business, you know, both of these. Go and work out in the long haul. It is a marathon. You've got to really deal with a lot of paradoxes. And work and life are just one such paradox. But you know, there's a short term versus long term. That's, people that you bring from outside who are leaders, but they come from the different biases. You know, some of them really are about short horizons, small horizons. The other ones are long horizons. And you need to really make them work together well with each other. I think the most important thing though is, people have to realize that at the end of the day, whether you build a small company or build a large company, it's equally hard. It just takes as much time to build a small company as it takes to build a large company. So my bias is go work on something that's big enough. and you might think that you don't have the technological prowess and the talent that you need, but If you don't have that ambition, you'll probably never build anything, uh, that really is, really market defining. I think we learned a lot about systems. I mean, I can tell you that as a CEO myself, I was personally involved in so many decisions and systems and systems to me is culture. Not just Slack, uh, you know, not just Gmail, not just GitHub, but there's a lot more to it than just that. You know, I mean, I was very involved in choosing our ERP system. For example, next week was a system of choice for us. And within a couple of years, SAP came knocking at the door saying, Hey, You will run out of steam with NetSuite because it was not built for anything above 100 million. And we have made it work. It's still running at 2. 4 billion annual revenue of Nutanix. NetSuite is still running because we actually put a system of engagement of Slack and bots on top of it that managers didn't have to go and log in. same thing with customer support. We said, we got to rebuild our own. We didn't use Zendesk as a service cloud. We built our own because data mattered, telemetry mattered, search mattered, you know, analytics mattered. If anything, I think, uh, one of the things that we've done at DevRev is to really think about, okay, so what's the operating system on top of which we'll build these next generation apps? Support is one of them. You know, we build product management, we build software development. We're going to build, uh, sales, uh, you know, relationship management with customers and buyers and sellers, all this stuff, but if the underlying operating system is not capable of delivering great search, great analytics, great workflows, I think you're building a cheaper, faster mousetrap, which the market is very little patience for. I mean, honestly, you can build a cheaper, faster, something for the SMB, but the moment you go to mid market in the enterprise, I think the bar is really, really high. So the one last thing I'll say on building. You know, the hard thing about hard things is how do you really go upmarket and we can delve deeper into the mid market and enterprise and the requirements that actually come from that because it probably is just as hard to build an SMB company as it means to build a mid market enterprise company.

Conor Bronsdon: 16:05

It's interesting. You bring up this idea of is your problem big enough and are you solving enough? Because I think there are plenty of folks who are listening, who are Saying, Hey, look, you know, I have a small app that are on the side. It makes me a bit of money and I'm happy with it. I don't, I don't need to scale it out, but there is such a difference in approach when you do want to scale to a Nutanix level, to a DevRef level. and so this like philosophical piece of it that you're kind of alluding to here is really important. And I know this also correlates to how you think about designing for scale and designing for AI driven scale. can you dive a bit more into how you have approached that from your product philosophy side of things and how you and your co founders are approaching scaling up to that next level?

Dheeraj: 16:53

Yeah, you know, scale is, a really interesting topic. and if you didn't weave architecture early on, then it's really hard to scale, you know, Manoj and I, you know, uh, I've known Manoj for 30 years. He was with me at Nutanix for about seven and a half years, uh, as well. And he and I, we talk a lot about this. Obviously, people is a part of it, process is a part of it, but product equally so. You know, SaaS 2. 0 as we call it, doesn't have to take sides between the enterprise and the SMB or the startup. You know, there's a false dichotomy between PLG and SLG. Uh, at the end of the day, even when you're doing sales led distribution, the fact that you have a zero touch, uh, product will be great because now in the large enterprise, yeah, the large enterprise, there's lots of small teams as well, who want to do shadow IT, who want to go and make, uh, autonomous decisions under the radar of what IT is. So thinking about an architecture that could equally scale for the freemium as much as it actually scales for the large companies was very important to us. And the way we build this cloud native solution was at the core of this. We said, look, uh, we have to have a budget for every small tenant. We can't spend more than tens of cents. On AWS for the small tenants, and that meant the way we carved out our multi tenant architecture for the tenant, whether it's a big tenant or a small tenant, we had to do it in a way that was not lackadaisical, because sometimes you can actually make these lazy decisions about architecture. They'd eventually go on to have you make this false choice that look, now my product is only built for the large segment of the market because the underlying cost of good ship, the cogs is so high that I can't, can't really go and cater to the other side of the market. And by the way, that segment is important because that's how the word of mouth grows. So if you want an efficient go to market over time, I mean. Look at Snowflake, look at AWS, look at Slack. All these companies, they had to build both ends of the market well. If anything, the product led motion, you know, delivers warm seats that eventually salespeople can go and call into and say, Hey, I heard that you've been using our product. It's been great. You've been spending all this money on a credit card. Let's go and show you the enterprise way as well, because now you're big enough for a salesperson to actually call you that architecture where you're really looking at the two ends of the spectrum and saying, Can I build a next generation company, which is a SaaS 2. 0 company that caters to both ends of the market? And that's why design matters in all of this. And the way you design your product is the way your go to market will actually be executed as well. I think, uh, in many ways, The last SaaS, they didn't have to worry about knowledge discovery. For example, search was not their problem because Google had not built something good enough for SaaS companies to deliver search. Analytics was not their problem. Like, no, well, now IT will take care of it. And, you know, building a great workflow engine was not their problem because integration was not their cup of tea. Integrating was IT's problem. The search was an IT's problem. Analytics was an IT's problem. And workflows, i. e. integration, were IT's problem. You know, and we just said, look, in the next 20 it better do more. It has to do more. Just like SAS 1. 0 did a lot for, you know, The market, you know, they had to left shift a lot of the complexity and said, look, IT is struggling with just building hardware, data centers, you know, upgrade cycles, software patching, all this stuff. We will do all that for you. And you know what? We'll deliver APIs that you can now use where you don't have to worry about the mundaneness of doing all this stuff. SAS 2. 0 has a very similar problem. It has to do more than what SAS 1. 0 has delivered, not the least of which is integrating everything together. Because now we have all these different SAS things, they're all fragments, they're not come, they've not come together. So we felt there was a big opportunity for really integrating one more time. And I think AI is a great catalyst for that. It's an amazing catalyst to bring things together because the only way you can go deliver on, you know, great chain of thought reasoning and an amazing self deflection and reasoning and all this stuff where people don't have to disturb other people to get simple answers. is if you actually brought the customer side of things, the product side of things, the people and work side of things, the users and their session side of things. So we make a case for a knowledge graph and without this knowledge graph, we don't feel there's real AI. And I think those are some of the lessons that I took away from my Nutanix experience. Look, I mean, as a CEO, I was acting like a knowledge graph. I was like, Hey, I think I know what this function leader does, that that function leader does. And you know, what's going on at the project in this department. And I would just try to be the, you know, the thing that integrates everything. So if anything, I think AI could be a great leveler where Any CEO can actually say, look, uh, I can use this co pilot that integrates all my functions into one.

Conor Bronsdon: 22:13

So I want to ask you about a concept that's been coming up a lot recently after Paul Graham wrote about it was this idea of founder mode. And you're kind of alluding to your philosophy around this. What do you think of this kind of push for a different mindset for founders? Do you think that maybe it's taking it too far?

Dheeraj: 22:28

You know, I think, the more things change, the more they remain the same. I look at Larry Ellison today and he has been in the founder mode for the last 40 years or more, right? I mean, uh, 1977 till now, he's been in a founder mode. I just realized as of yesterday that he's richer than Mark Zuckerberg. Because of. How he's always tried to come from behind and win markets. So, I don't think much has changed. It's not a new definition. Of course, when Paul Graham writes something, it becomes news. But at the end of the day, I think, the fact that you have to be Extremely, ownership mindset, you know, the founder mentality, the founder's mentality is an idea that more and more people in a company need to have, you know, Bain Company did a really good, uh, work on this almost 10 years ago. It's, it's a, it's an 18 minute video on YouTube, the founder's mentality. And it talks about how, as you try to let go of the company, you know, You fail and then how need to get back to having those people who actually have the founders mode as well. So I think this discussion has been going on for a while and I don't think much has changed in that. I feel like, this is where I struggle every day. Uh, it was like, have we found the next inner circle? And have you found the next three to five to seven people who are the next set of founders of this company? So technically they might

Conor Bronsdon: 23:49

mentioned that raising up leadership

Dheeraj: 23:51

Yeah. No. And I think, and that's the hardest piece. And remember if it goes from two of us, like Manoj and I to the next seven, 10 of us, all of a sudden the company probably has four to five times. A higher chance of scaling and succeeding as well, you know,

Conor Bronsdon: 24:06

Well, so let me ask you really directly here then. So like, how do you and Minaj go about getting those folks to buy into that level and become these kind of instruments of, like, almost co founders themselves?

Dheeraj: 24:16

a big part of it is how they feel like it's their company, which is about, uh, you know, I mean, a lot of it is about letting go of how you are always right, because you're a founder. Right. I mean, the mother has to go, let go of a child because the teachers know a little bit more. Right. And over time, maybe their, girlfriend and their fiance knows a little bit more and so on. So I think, you know, this is Right. Thanks. Something that is extremely natural done well or not, because tensions get created along the way, you know, parents fight with teachers about how little teachers know about things and so on. And I think this is the core of it. I have a Pareto on this, like 80 percent of the stuff that you let them make mistakes. Don't try to Be right in everything on that 80 because they also have to learn from their mistakes, you know? And under 20, you need to have another 80 20, which is like maybe 16% of that hundred. You go and argue. You spend time, you negotiate, you try to tell them why. What you are saying. Probably also needs to have weight, but eventually that rumble will create good decisions, and it's only in the last 4%, which is the 20 of the 20. That you really throw your badge and say, look, I'm telling you, just follow the gut. I have a gut in these things. And what founders fail to realize is that if they don't have this kind of a mindset, then everything is about throwing a badge. And look, because I'm a founder, I can tell you what is right. I think that primogeniture Uh, this thing that look, I seeded this, I created this. Therefore, I know the most is what really results in unscalable companies. And if you can actually create this thing that look, people can come from behind and still win and still feel like they are the owner, that's the way you also create ownership in these people. Like, you know, in my last company, we had, somebody who had never done sales before, and he became the president of the company over the seven, eight years. He also went on to become a CEO of ThoughtSpot. from where he started as a senior director to where he ended, I had to really build a second founder in the company because I'd lost two of my founders between 2010 and 2012. and I think it mattered that he actually said that, look, I can make a lot of these decisions now, uh, and so on. So how you nurture people to become founders again, even though technically they're not, is at the core of this. And, you know, those. Things that, Brené Brown talks about being vulnerable, you know, what does it mean to actually show that vulnerability that look, uh, let them go make some mistakes along the way is probably the core of this too.

Conor Bronsdon: 27:00

I've got a couple of her books over my bookshelf behind me, so

Dheeraj: 27:02

she's amazing. She's amazing. And, you know, she talks about some acronyms that I keep in my head about, like, WGLL, what good looks like., And WGLL can be applied to anything in life, you know, people, process, product, you know, i. e. design and, you know, all sorts of things, you know, what good looks like with respect to relationships, what good looks like with respect to processes and product, just amazing,

Conor Bronsdon: 27:25

I've maybe been under applying that one. I haven't really thought about it. Uh, in that perspective. So let's focus in on that product perspective a bit more because we've talked a lot about people and the importance there. And obviously that's integral to scaling a company. But as you pointed out earlier, there is not this clear dichotomy between Product led growth and, you know, sales led growth, which is really in a lot of ways, people led growth. Like both of these concepts are needed to have a company succeed at the highest levels and they both, uh, work closely together. I've read some work by you on Minaj. Uh, Minaj wrote earlier this year, a great blog about designing for moments of serendipity within the product and, and enabling those, uh, letting users make these like deep connections in their brain through like happy magic they experience. And then obviously you've talked already today Built in AI and building for that scale versus kind of the bolt on AI that some other companies are experiencing. What's your perspective on how you think through AI driven design and how DevRef is designing for the future?

Dheeraj: 28:30

know, when TikTok came out, it took the world by storm and the reason was simple. It was not syntactic, it was semantic. It was really understanding what your tastes look like. I think the fact that it was a great recommendation engine without you having to even ask it a question was at the core of its virality. I mean, obviously, everybody else copied after that. Now, whether or not their neural networks was as powerful as TikTok, I think time will tell, but I think that's the core of software going forward is like, can it really be serendipitous? Can it really go and be the best recommendation engine? Can it know more about me than I think I do? About myself, uh, many of these things surface up from the subconscious and softer. I mean, and it does get creepy sometimes because it's trying to collect everything about you, but you know, if you keep that privacy guard aside, especially if you're willing to actually let it know everything about you. I think in B2B, there's a big opportunity for serendipity, for recommendations, for letting it tell you, like, you know, Dealing with analytics for now and, uh, analytics has been a great struggle, you know, in terms of what insights mean. And up until now, there have been big teams of people who actually do data pipelining, massaging work, then warehousing work, which is going and summarizing and aggregating and putting jobs around it and finally going creating visualization, i. e. dashboards and reports on top of it. I mean, millions are spent in every, even a fairly small company, millions are spent every year on data alone. But if you think about at the core of it, it really is a search problem. You know, many people are building these widgets and visualizations, things like that. And many a time, I want to know what others have seen. After I asked one question, maybe the next three questions have to be serendipitous, where the recommender actually tells me, by the way, this is the next thing that they've asked, or this is the next drill down that they actually looked at. People who did this also did that, also made Google a little bit more You know, usable. In fact, if anything, that's the level of serendipity that we actually crave from all search engines, you know, and analytics is now a search problem. So, I think what we had to do was to think a lot about design, uh, of these kind of systems, and we're still scratching the surface, but I can tell you that, you know, at the core of what B2B will look like in the next 20 years, it has to be more consumer grade, it has to be more TikTok like. it has to be design worthy and real time. You know, I mean, in the last 10 years, the best that we've done with B2B is collaboration with Slack. That's the best. It's the epitome of design is, is collaboration because it has transcended the departmental boundary that's like, well, we can get anybody across all departments to come and work together. If you don't bring that to business software, we have failed. Because the problem with Slack and Teams is that they're very unstructured. All they are is a system of engagement. But you've got to bring that into your tickets, into your issues, into your enhancements, into your conversations, into the way these chatbots are actually working between users and the companies and so on. There's a lot of collaboration. There's a lot of Slack like magic that needs to be brought to business software. And we feel like that's the missing piece that will take, you know, this Right now, the way we're talking about, uh, you know, software and work and people, like, are they willing to work as hard and things like that? We've got to make it delightful and, and both AI and design can actually bring that element of delight to business software.

Conor Bronsdon: 32:18

I love that idea of delight in business software because it's something we often don't think about enough. Another concept that's really important, which you mentioned there, was privacy. And my understanding is that DevRev is also applying kind of the principle of least privilege, which is a cybersecurity concept that limits a user's access to the minimum level of permissions needed to complete a task. to help protect that privacy and make the, uh, what you're doing more secure. Is that correct?

Dheeraj: 32:44

Yeah. And by the way, these are lessons from Nutanix too, because we kind of took longer to actually think hard about, uh, principle of least privilege. Because sometimes when you're building simple, easy to use products, you tend to forget that at some level in the enterprise, They will have confidentiality, privacy, integrity, all sorts of things to actually go and, and think hard about, because it's not one company, it's small companies and larger company, you know, and every little department or group actually wants its autonomy and its, privacy as well. So we had to really DevRev. To really think hard about security, we think we've built something that's even better than the best authorization engine that's out there, you know, Google proposed Zanzibar, which people think is extremely fine grained and, and so on. And again, the moment you start thinking about fine grained security, how do you make it easy to use is as much of a challenge and opportunity as anything else. So yeah, I think. Privacy has played a big role. Security has played a big role. We're also thinking about multi region because, look, the world is de globalizing. And in a, in a more de globalized world, you have to really think about the laws that land more now than five, even five years ago when cloud was actually coming out. They probably didn't have to worry about those things. And now every country worth its salt is thinking about data residency. And this is the place where a lot of SaaS 1. 0 is struggling. You know, look at Zendesk and Service Cloud. They can't go punch a cloud every 15 days because they're built with private data centers in mind, or they're built with large AWS thinking in mind, as opposed to, can you miniaturize this and bring it to a country? Like, I don't know, it could even be like, I have a data center in Saudi Arabia and another region in Israel, or wherever the hyperscalers are, uh, one in Southern Japan, we can do that. And SAS 2. 0 has to go and miniaturize itself. Because that's what the new cloud's definition really is.

Conor Bronsdon: 34:45

You brought up a couple of great things that I want to dive into. but let's start with that decouple globalization piece, because to your point, we're seeing that happen around data and regulations. We're seeing regions and countries really get serious about keeping data in their region, uh, supply chains, look at the chips act here in the United States, uh, and some of the concerns there around hardware. But it's not happening with people. In fact, the opposite effect has happened, particularly since 2020, as everyone has kind of pushed online and remote. We're seeing global recruitment. We're seeing more and more hiring across different regions. Even companies that are doing return to office are trying to leverage this massive talent pool globally. How do you see these trends

Dheeraj: 35:32

well, more than conflicting, I feel like, work and life are weaving in together better than we had done before. Now, it started with video, and the fact that video is now a commodity. was not a commodity before 2020. WebEx, most of us didn't think, we didn't, you know, actually think of video as the first last thing with WebEx. But in the last five, seven years, with the advent of public clouds and big backhaul networks and, uh, you know, client side caching, HTML5, there's a lot that's happened with video in the enterprise. But it has to go deeper. And this is where I really believe the knowledge graph needs to be, uh, coming in because now the hallways are gone. You know, you're not gonna bump into people in the hallways. So what's the digital version of it? The digital version of it is what we call the knowledge graph. The customer, the product, the people, the work, the user, the sessions and activity. We need to bring this new sort of Temple, the building where everybody actually lives. And Slack was that, has become that in many ways around collaboration and channels and things like that. But I think we need to think about AI as the new sort of place where people It's the new public, say, you know, central public square or something. You think of the place where people congregate. It's no more the, I would say, the atrium of a building. You know, there's no more atriums, uh, atrias, atria where you can actually get people to congregate. Uh, a lot of these things need to happen digitally. Now, there is definitely this move towards, you know, the big cities have become too expensive. We need to really do this, in a way that can be best to both worlds. But along the way, I think there is a little bit of a schism between, uh, the West and the East, I would say, you know. I would say that emerging economies are showing more hunger, uh, in many of these things, you know, saying, look, we have nothing. And we'd like to do so much more. the experience in the West and the, Youth these, we need to mix these things together as well, you know, I think, and, the college folks here, uh, coming from, let's say American colleges, they have to also basically say, look, the world is flat. We need to really go compete, uh, with, you know, this amount of hunger that's come to develop because video is everywhere and collaboration is everywhere. And the knowledge graph can now be spread across. we will actually see a lot of this stuff where the world has actually become flatter. I mean, that's the, one of the reasons why Debra actually had, you know, seven offices in the first three years. I could never have imagined this at Nutanix. It took us four years to even start a second center of excellence out of Bangalore. And this time around on day one, we actually had to think about Slovenia and India. And then day two we had to think about Argentina. talent is everywhere right now. And, um, we have to think about this as a great leveler. And at the same time, the fact that it's become as hard as it was in the 80s, like, you know, 80s were not easy for tech jobs here in the U S uh, because software is barely coming out of, uh, it's hardware roots, and, uh, in the nineties, early nineties, I think until the internet bubble, I think. Software jobs are not easy to come. And, uh, I think we're back to that where we, again, by the way, hardware is very important now. GPUs, as you know, is the new revolution. So really to get competitive again is like going back to the 80s.

Conor Bronsdon: 38:58

It's interesting you bring that up because I think one of the undervalued changes that happened in the 80s was also the investment in particular computer science programs in certain cities, uh, whether we're thinking about like Stanford and Berkeley or we're thinking about, you know, University of Washington, Seattle and what Microsoft invested into that program and Paul Allen in particular. And so we've seen obviously, okay, like we've kind of begun to understand that, you know, To have a great startup community, a city that has a ton of talent, you need that education level. We've seen the internet start to democratize some of that education and bring it global. And we've also seen other cities and countries across the world make those into educational investments. And now to your point, we're seeing this next layer of, okay, Great, we have this huge cloud push and now it's we're really going all in on GPUs because it's going to be fueling this next layer of AI and the capital investments we're seeing from particularly major players, you know, the Apples, the Metas, the Microsofts, the Amazons, the world. Uh, it's really, really exciting. Crazy to see the scale. And I'm very interested to see whether open source GPU power becomes something that actually really moves forward and whether like the llama approach or something else, really gets some precedence or whether it stays very much in house and people are looking for this monopolistic win,

Dheeraj: 40:10

Well, I mean, you know this as much as anyone else that, centralized architectures eventually decentralized. And the big reason is access and economics. when I say access, I mean, can I run it on my laptop? You know, that's how open source came about. Like, well, I don't want to pay for a cloud provider's ability to really do this. And latency matters too, because if I'm on a free version of OpenAI and that takes 15 seconds for me to Get back an answer because I'm on this unprioritized queue. Developers will find better ways to do that. And they said, look, I can just use a Lama 8 billion model. And by the way, that's really good. And I can fine tune it and so on and so forth. So the years out architecture is well understood. It will be open source. Running on laptops, uh, because the laptops will have a lot of good GPU as well, you know. Uh, now, in the next 5 years, uh, maybe next 5 10 years, I would say that, uh, open source will copy a lot as well. Like, the good thing here is that, Meta has a lot of these things going on in their consumer cloud too. So they're learning a lot about languages, they understand a lot of languages. So I'm actually rooting for open source at some point, uh, you know, to become as ubiquitous because I personally have believed a lot in edge computing. If anything, the way we've done our analytics product is and the way we've kept it democratized Just as accessible to a freemium client of ours, as it is for a large enterprise paying us millions of dollars, is by using their laptops. Like, hey, we got to use the laptop as much as possible because I can't burn so much money into centralized data warehouses, for example, to do analytics. We've got to really figure out the way Excel popularized or the way Tableau popularized where the data was on the laptop. And now people could do things so much faster rather than going back with a round trip to a server and that to an expensive server like a centralized data warehouse. So our analytics architecture, for example, I mean, there's nobody that even comes close. Um, the way we ship data from, you know, uh, these different systems record, the way we bring it to a centralized data warehouse, the way we crunch the numbers and summarize and aggregate, You know, bring it back into AWS and then bring it to a laptop, you know, in a highly secure way and then make people do things on the laptop, I think is the essence of, uh, and not just laptop, but eventually mobile phones to, uh, I mean, voice to text, text to voice, all this stuff will actually happen on the edge device and we've got to make, that's the only way to really democratize AI.

Conor Bronsdon: 42:51

I mean, it's interesting to bring this up because, uh, I mean, we're recording this a few weeks before the episode is going to come out, but open AI just in the last, you know, 24 hours has announced their new open AI, uh, Oh one preview based off of Strawberry where they're actually using. Essentially slower thinking from the AI to get to a deeper answer. So it's going to be fascinating to see. This kind of these kind of two trends to your point where it's like, Hey, we want to have deeper models that can take their time and get to like a more well thought out intuitive answer that has deeper understanding behind it and gives us a more extensive approach. And then we're going to see. Um, you know, the digital employees, the agents, so does, uh, the agentic AI approach where it's like, Hey, like, I need this digital employee to do things regularly, regularly. And then we're also going to see the kind of, I'll call it the co pilot approach of like, Hey, let me leverage this thing. And it needs to be extremely fast and low latency and, uh, to your point, available locally. So those trends and how they interact is going to be fascinating to see over the next couple of

Dheeraj: 43:47

Yeah. And look, the idea of slow thinking, it doesn't have to be a pause of 15 seconds. I mean, even the most deep experts, they don't pause for 15 seconds for the next answer. So we should probably, hopefully try to aim for two second pause. Like, Hey, let me think about this for two seconds.

Conor Bronsdon: 44:06

I think we'll get there. may take some

Dheeraj: 44:08

and this is where I think open source can be very helpful as well, because not everybody can afford the tier one enterprise licensing, open AI to really go build that two second latency. And look for most startups, latency matters. Actually for all startups, latency must matter. Because if the end user experiences that I don't know what's going on, and 15 seconds is just too much. Google, Google tried for a 300 millisecond response and that's why they became so popular. I think if things are in the realm of a few seconds, we're already violating the principles of attention spans.

Conor Bronsdon: 44:45

Well, I could honestly talk the future of technology and kind of hear from your insights, uh, for another hour here, but I know we're, we're getting close to the end of this episode and I want to make sure, uh, to ask you a couple of key questions here at the end, because it sounds like you've got some exciting news around DevRev's leadership in AI to share with us, Dheeraj. Can you tell us a bit about, What that news is.

Dheeraj: 45:05

Well, uh, the big news that, I'm very proud of is to really Get, a leader for the company at a board level, somebody, the first, uh, external board member is somebody from the world of design. His name is George Pecheneg, and, uh, he is one of the special people who's worked a lot in consumer companies, and also in enterprise companies like Microsoft, and then finally, At New York Times, where he was really spreading a lot of the design experience of New York Times. I mean, I am in awe of the New York Times. I always wanted to really get in touch with somebody who had done so much work, uh, especially look at New York Times the last 10 years and the way they've really redefined themselves with. Design and now, you know, games and one of the most popular engagement, uh, sort of apps out there, New York Times apps. So I really want to, wanted to actually think about how we started with design because it's a counterfoil, you know, and I've always thought that when the world is talking AI, AI, AI, AI, AI, we have to think about the opposite of AI, which is UI and UX. That if you can bring this well, then you can make AI invisible. Uh, it's like the way Warren Buffett talks about it. When the world is greedy, you should be fearful. When the world is fearful, you should be greedy. So in this day of a lot of AI washing, we figured we had to do a really good job of design. And hence, uh, George Petchenik. I think it's going to really be a perfect foil for a lot of the, you know, deep technical thinking that a lot of our developers bring, but also a lot of what our design team will need to really be a first one equals four. We had to really think hard about this for the last three and a half years, because up until now, uh, we didn't have a board. We, we actually had just two of ours, uh, Manoj and I on the board. And now we have an external board member who really, You know, uh, is carrying the torch of, uh, how to make AI actionable and useful.

Conor Bronsdon: 46:57

And this speaks to your philosophy of investing in people and building out excellence within your team and within your advisor set too. So I see how this aligns to kind of the lessons from your previous journey and also how you're thinking about future product philosophy, which my understanding is, uh, DevRev has an exciting upcoming conference with some new platform announcements that's coming up here about a week after this episode

Dheeraj: 47:20

Oh, absolutely. I mean, uh, I'm gonna, uh, spill the beans a little bit right now, but at the core of this is to not just, you know, this idea of change, because AI, people are, Afraid to say this is bolt on or the built in and we're like going and making case for like when everybody thinks it's a bolt on you have to really build it in. But building it in means that you have to replace things. I mean there's no easy ways to actually go, you know, transplant a brain from the outside to make one body work. In many ways, you have to think about this whole thing holistically. So while we will talk about what AI design can do to a business, but customer support at large, product management, customer support being next to each other, but also what do agents look like? What does incident management look like? You know, what does, what does insights mean in the enterprise beyond search? What does search itself mean? And how can you commoditize search to the point where you don't have to pay hundreds of thousands of dollars to actually make enterprise search work? I think last but not the least to talk about case study, you know, how did we do multi million dollar foundations out of Zendesk and Salesforce service cloud? And in the future, how do we really make people think about the enterprise worthiness of next generation platforms, how being cloud native, being AI native is the only way for people to really step away from the treadmill that they've been on, because most of these technologies. Lines of business leaders, they've been on a treadmill and they're all hoping for, uh, what will change and the only way you can actually bring change is when you think hard about not just slapping something on like, uh, I slapped on AI and now I can get a promotion and instead of think hard about What do I need to really change?

Conor Bronsdon: 49:02

I love it. Uh, well, Dheeraj, I'll make sure that our listeners, uh, are paying attention here and say, like, check out the upcoming conference effortless by DevRev on October 17th and Santa Clara for these new product announcements and so much more from Manoj and Dheeraj and the whole team at DevRev, uh, we'll link the event in the episode description, or you can find it at devrev. ai slash events. Dheeraj, thank you so much for this fantastic conversation. I think our listeners will benefit a ton from your insights. Hopefully it's going to inspire some folks in our audience to, to Take the next step as a founder or to grow their careers, uh, or, you know, become a co founder in name and spirit. so thank you so much again for joining us on the show. And for those tuning in, you can find this whole conversation on YouTube, uh, across all your podcasting apps. And don't forget to leave us a rating when you have a chance. Thanks Dheeraj.

Dheeraj: 49:47

Thank you.