Rajeev Singh: AI Agents Are the New Users

[00:00:00]If they bought the company for $8 billion, they're looking for a return.

[00:00:02]That return wants to be, you know, in the neighborhood of $20 billion. We said uh MCP is the future. Ultimately, AI will be the new UI and all of the functionality of Smart Sheet will be available conversationally. The story for enterprise software companies is fundamentally I think really three core areas Sasha.

[00:00:22]The first is do you have context? Does your data provide value in a way that makes agents and human beings smarter in their capacity to do their work?

[00:00:32]>> Hey everyone, before we are starting with the episode, I need to ask you for a small favor. Every week we are doing a lot of research to get you the best insights. Please support us, hit the follow button, share the podcast with your friends. Thank you very much. And now, let's start with the episode.

[00:00:56]Hello everyone. Today I have Rajie Singh, the CEO of Smart Sheets with me today. Great to have you here.

[00:01:04]>> Thanks Sasha. Great to be here.

[00:01:06]>> So I think you once said if this transformation moment weren't happening, I probably wouldn't be here. Does it mean like the AI wave is the only reason you took the job as CEO? Uh >> I took the job nine months ago. It's not the only reason. It's not the only reason, Sasha, but it's a big reason.

[00:01:25]It's a big reason. Uh I was the co-founder of a company called Concur a long time ago. Uh I guess now almost 30 plus years ago and then was uh the CEO of a startup called accolade after that.

[00:01:39]Both companies were lucky enough to have some success. We took them public. We built them. Uh many people have asked me in the past you know why are you back doing something new at this stage of your career and I thought uh this moment of technology transformation AI was just a wave that if you if you love building companies and if you love technology I couldn't miss and so it's a big part of the story now the other part of the story is Smart Sheets a company with 100,000 customers and you know billions in revenues and a huge opportunity to actually take adv advantage of the AI wave. But yeah, uh that's a that's a correct quote. I did say it and I meant it. This is the most fun time ever to be building software and I'm glad to be doing it.

[00:02:25]>> You stud you studied engineering, right?

[00:02:28]>> That's right.

[00:02:28]>> Back in the days, did the the engineer ever left the room or is it still with you at every meeting you're having?

[00:02:36]>> Uh it's still there. It's still there. I think engineering school gives you whether you went you know whether you went to engineering school or you worked as an engineer building software even if you didn't go to engineering school it gives you a structural mode for how to solve problems uh you know what what you know starting with first principles and then working your way through a framework and so that never changes now my engineers would tell you at every company I've ever been at uh that I'm not really a software engineer I'm I'm a mechanical engineer and uh they know how to build software and I'm more of a product person in that regard. Like I consider myself way more of a product management ilk than a engineering ilk as it relates to software companies. But no, the engineer never leaves the room.

[00:03:22]I I I I feel like it's passed down to me from generations of engineers in my family. Is it especially now the job at the CEO of Smart Sheets that you can benefit from these engineering background especially when it comes to project management and these kind of things?

[00:03:39]>> 100% Sasha 100%. I think partly Smart Sheet is built to get work done uh which is fundamentally you know engineering at its core is how do we move the world forward uh by building things. I think the second part of that story is like if you look at all the software look at all the big software companies right now founders are coming back to their companies uh because the product needs to be reinvented from the core. This isn't about uh this isn't a moment of how do I operationalize things. It's a moment of how do I reinvent things and that is at its core a product and engineering task. And so I think that's it. It it is a moment for engineers and product people to rise to the forefront, right? You're starting companies all the time. You get it.

[00:04:27]>> Engineers are going faster than they ever have. Product people are rushing to catch up to engineers, which has never been that way.

[00:04:34]>> Uh that makes it a really exciting moment to be building.

[00:04:39]>> And I think it's the struggle right now is really to catch up with all the information, right? Like I mean I see it all the time. the the the developers are building and building and building and the thing is how do you protocol what you have built and how do you all put it together in a single output?

[00:04:59]>> Yeah. Yeah. I mean Sasha I'm on two startup boards right now. One where I'm the executive chair and co-founder and these companies that are starting from scratch are rewriting their code base full stop. rewriting their codebase three every 3 to 6 months like because the changes to the frontier models are so significant that they think oh you know what the cost of writing code is so much lower than it used to be I'm just going to throw away what I wrote before and rewrite it uh I mean that's just mind-boggling to somebody who grew up in software for the last 30 years >> yeah absolutely so maybe we can talk a little bit about smart sheets because I have actually a background in project management so I did a lot project management all all across Germany and to my days we worked with Excel maybe Microsoft projects >> what makes smart sheets different like today like how does it work with all the AI tools like what's the benefit >> so smart sheet at its core is a tool to get work done across systems of record so you need to pull data from workday from Salesforce you know systems of record like workday and salesforce are great in their departments but not great to get work done crossf functionally.

[00:06:17]Smart sheet solves that problem by linking together those systems of record and creating complex workflows and automations to get work done. That's the bottom line of what the software does.

[00:06:28]Now, traditionally the software was built for a PM, you know, a project manager uh or a program manager like you mentioned Sasha who built their own automations, rules and custom capabilities or workflows inside of Smart Sheet. Uh AI has changed the capacity for actually leveraging the tool. And you know, one of the first bets we made uh when I joined the company eight or nine months ago was ultimately fully embracing MCP. We said uh MCP is the future. Ultimately AI will be the new UI and all of the functionality of Smart Sheet will be available conversationally. And today we're delivering it via Claude, via Copilot, via Gemini, and via ChatgBT. So it's not just for project managers anymore. If if you can conversationally communicate, hey, I want to go get this.

[00:07:20]I want to go get my resource data from workday. I want to go get my customer data from Salesforce and I want to go get my invoicing data from Netswuite all via MCP and I want Smart Sheet to run a project in order to achieve the following tasks. You can do that conversationally and you you actually never have to open our software to get it done.

[00:07:40]How do you see like the whole project management um is moving over the next five years? I mean, if even if it's even hard for me to catch up with all the LLMs and I'm like really >> Yeah.

[00:07:53]>> Yeah.

[00:07:54]>> very deep in the industry.

[00:07:56]>> Yeah.

[00:07:57]>> I I think what's going to happen one, the PMO function will get 10 times more powerful because it'll have 10 times the capabilities, more tools. things like smart sheet will be available to it in a very simple way. Uh and smart sheets really deep meaning uh we differentiate ourselves by being an enterprisegrade security auditability rollback capabilities structured data sets all that deliver context to an AI engine so it can be smarter about the decisions it makes so your agents can be smarter about making decisions. Uh that part of the story will make PMs more uh incredibly powerful. I think the flip of that is as well, uh, you don't need to be a PM anymore to use the software. And so, you know, somebody in marketing can say, I have a new pricing plan I want to roll out. They don't have to go to the PMO office to roll it out. They can use Smart Sheet to go solve that problem on their own. Uh, agents will become incredibly powerful PMs. And so, PMs won't have to learn the LLM. PMs are going to have to learn the tools, and they're going to have to empower agents to go deliver those tools.

[00:09:03]Are you also with Smart Sheets facilitating these agents?

[00:09:07]>> 100%. 100%. So for you know if the first bet you placed as MCP, you've placed a bet on agents, right? You're saying humans have been my customer forever.

[00:09:17]Humans will continue to be my customer, but I have to build a software so that agents can use it as well. And agents are going to use MCP to go to to deal with my software. So Claude Claude Co-work or agents you build on your own.

[00:09:30]We don't care where the agents come from. We're building agents on behalf of our customers. But we do fundamentally believe that we can deliver a a cat a whole catalog of sub agents. You want to do a risk analysis on your project, I've got one. I've got a risk analysis sub agent for you. You want to do an executive summary, I've got a risk I've got an execum uh uh sub agent for you.

[00:09:54]These sub aents will be called by your orchestration agent. Either we can build that for you or you build it for yourself. The whole idea though Sasha I think is is open systems no lock in give customers all the access to value. Not every software vendor is making that choice but that's the choice we're making.

[00:10:13]>> So you as you said you are open to all models right? Is it Microsoft? Is it Google? Is it Anthropic? Have you ever thought about building your own LLM?

[00:10:22]>> Uh we've thought about we certainly experimented with open source LLMs. Uh and right now we're kind of optimizing Sasha for speed. Uh we want to deliver more and more software as fast as we can because we see the market moving really quickly. We want to be the leader. We've always been the leader. We want to continue to be. Uh but we are experimenting with open source LLMs where we think uh price to performance is incredibly uh is incredibly good. I think the other thing we'll do is as we find deeper modes around our context, uh we might choose to use open source around those those particular context areas that we think are most valuable.

[00:11:05]But right now to be honest uh it's a moment for speed in my mind and we're running as hard as we can. So that means we're leveraging we're kind of using some mix of anthropic and uh in some cases Amazon models where we're looking for lower cost.

[00:11:21]It probably also has a lot to do with risk mitig mitigation, right? Because I had it once with a new model from anthropic coming out. None of my previous workflows worked anymore because they changed something how the LLM responded and this change made all my workflows not working. Like yeah, >> how do you deal with that?

[00:11:45]uh you have to be we have to we have to be incredibly thoughtful about embracing new models inside the software and so and to be candid with you Sasha you don't need Fable or Sonet 4.8 eight or you know for everything you do and so many of your tasks are simple enough >> that you can be using Haiku or you can be using uh you know a sort a simpler a simpler version of a model and if that's the case and you're achieving your work you don't need to upgrade every model every time and so each task should be evaluated from a price to performance basis and you shouldn't be upgrading all the time you can't if you're building enterprise software you know we've got uh four million active users a day we can't we we can't break uh and so the software has to work every day because businesses run on it.

[00:12:35]>> Do you operate on a global scale with smart sheets like where are your core markets? What is the core industries you are having or are you all across the board?

[00:12:45]>> Uh we work all across the board from an industry perspective but there are some industries we're particularly successful in but first let me start with the geography question.

[00:12:53]>> Yeah.

[00:12:53]>> Uh the company was founded in the United States in uh in Seattle Washington. Uh today we do business our core markets Japan, Australia, UK, Germany, uh France, uh Scandinavia which I'll call uh a market uh and uh we intend to keep growing. Europe is going to be one of our biggest growth markets and uh Germany specifically for the purposes of our conversation today, Germany specifically is one of the markets we hope to grow in a a significant amount over the course of the next five years.

[00:13:29]Okay.

[00:13:31]And um when it comes to the industries, I'm just thinking about like what is the the the main industry that your company is operating in? Is it is it energy? Is it automotive? Like what's the core industries? you know industries with uh regulation structure uh >> you know so think about to some of our biggest most significant use cases pharmaceuticals life life sciences healthcare uh some of the biggest life sciences companies or healthcare companies uh some of the biggest pharmaceutical companies in the world use our product to manage their entire new drug discovery process from end to end uh we also have a very strong foothold in technology. Uh in Germany, uh buyer's a customer in the pharmaceutical space. Uh Adidas is a customer and runs a lot of their uh sort of endtoend workflows with us. But if you were to think about core focus areas, think government, think manufacturing, think life sciences and healthcare, uh and think technology.

[00:14:37]Those are the those are some of the the areas if you're to look at the billion plus of revenues we do. uh there's you know north of 30 or 40% of the revenues come from those sectors.

[00:14:48]>> So you joined Smart Sheet shortly after it was taken private again right for I think it was 8.4 billion.

[00:14:55]>> So is there anything Blackstone or Vista tell you what you need to tackle?

[00:15:01]>> Uh you know I think uh Blackstone and Vista are two really smart software investors. They're two really smart investors. And so you know uh what are smart investors talking about right now?

[00:15:12]Same thing you and I are talking about Sasha. Uh you have to make your your company relevant in the world of AI. Uh you know all of us have been reading it's a little overblown but some of us all of us have been reading about the end of enterprise software. So enterprise software is dead. It's the it's the age of enterprise AI companies.

[00:15:31]And you know, I think if you're to if you're to uh wipe aside the hyperbole of those statements, there's some truth to it. Uh companies that don't embrace not just plug in or sort of uh you know, think about a pending AI, but instead fundamentally rethink their software in the world of AI are dead. And that's very clearly one of the focus areas from Blackstone and Vista. Obviously, the other part of that story is how do you do that in a way that efficiently drives value? If if they bought the company for $8 billion, they're looking for a return. That return wants to be, you know, in the neighborhood of $20 billion. Uh how do you how do you drive profitable growth moving forward and I think we're uh we're making great progress in that regard.

[00:16:17]>> How do you I mean it was a a great catch like how do you really see the risk for software as a service provider in general? because I mean [sighs] every time a new model gets the gets launched um many many stocks are plummeting because there is a risk that they're getting obsolete. Um how do you see the overall market >> when co-work launched right the the market lost $300 billion of value.

[00:16:42]>> Yeah. Yeah. It was insane.

[00:16:44]>> It was nuts. Uh and yet if you use I I mean I am a vorac I have a voracious appetite for using AI models. I use them constantly and I try all of them from Grock all the way through to open AAI. We use anthropic internally at at accolade or excuse me at Smart Sheet. The the the the story for enterprise software companies is fundamentally I think really three core areas Sasha.

[00:17:14]The first is, do you have context? Does your data provide value in a way that makes agents and human beings smarter in their capacity to do their work? If your workflows or if the work you do is incredibly simple and easily accessible via LLMs, if that data set is easily accessible, meaning easy to accumulate and deliver by startups and LLM, your business has a problem. Now, uh I think fortunately for Smart Sheet, we have an incredibly unique data set and we can deliver that context. The second part of that story is uh I think what's referred to as proximity to intent. Are you close to the work getting done? When people think of work, do they think of you and do they choose your tool to go do that work? Again, uh Smart Sheet checks that box. And then the last is uh does your data set inform making work get smarter and smarter? workflow orchestration, but workflow intelligence. If those three things are true, you have an opportunity to build a business. If any of those three things are not true, you better get to work on solving that problem. And there are there are plenty of enterprise software companies. I think specifically, Sasha, the ones I think have work to do right now are the ones who are very smoke focused on small business. Because when you're focused on small business, you're typically focused on really simple workflows that can be deployed really easily. and for low cost. That sounds like a recipe for an LLM. Like that's where a co-work is gonna going to thrive. Complex, structured, regulated workflows uh a little bit harder. I use LLM constantly as I started with this this answer with and you know what uh even today cloud co-work disappoints when it when you're trying to do contextually aware hard work. And so uh so you know that's that's where enterprise software companies are supposed to play.

[00:19:07]>> What's your what's your favorite LM right now?

[00:19:12]>> Uh boy I I probably shouldn't answer that because they're all customers and uh all these guys are customers.

[00:19:17]>> No they are all great.

[00:19:19]>> Uh they're they're all fantastic. I am I am particularly focused on uh Gemini and Anthropic right now. uh both that I think uh anthropic more for my product and tech use cases. I I'm I'm in cloud code uh looking at our lakehouse data of our customer data every day or often. Uh and I'm doing that via AWS bedrock but cloud code. Uh so for product and tech use cases I'm very very cloud focused. Uh for other business research uh and thought partnering I'm using Gemini more and more.

[00:20:02]>> Yeah. Yeah. I also try to be up to date with all three or four depending on how you count who who is a frontier lab but yeah I mean they are kind of similar and even if not they will bring out the next model which is kind of similar to their competition but >> yes >> for me it's always the question I mean especially now with Fable um the Fable launch when it gets banned by the US government how fast will we see that it's generally gets restricted that at some point we have just like this is the level of LLM you get as a as the as the general public everything else is behind closed doors do you see that coming >> uh well mythos was obviously like this first giant test case of what's going to happen and you know I think there's speculation about how much marketing that was versus how much actual reality that was >> uh I think governments will get more and more involved in approving new models.

[00:21:03]That will be true. That just has to be true. Uh if you look at the number of vulnerabilities that were exposed by methos, but that will continue to be exposed by new models and uh financial institutions. So, fintech particularly impacted but highly regulated industry impacted. There has to be some mode Sasha there has to be some mode of protecting the public interest as new models outpace us. And I can't imagine anybody's can do that but the government. And so I I don't believe that most governments will be incented to restrict forever access to these models, but I think they will slow down the pace of innovation for the good of the public safety.

[00:21:44]Where do you think how where do we stand right now when it comes to policies and regulation around AI? I mean Europe is already having um killing kind of innovation with the European AI AI act. I think uh the US is is more open than Europe. But I'm not sure like which way is the right one.

[00:22:07]>> Yeah, it's a tough one. Europe has always been more conservative as relates to data privacy, as it relates to uh model regulation now uh and will always be that I shouldn't say always. Who knows what the future is, but Europe has always been and I think will likely continue to be more uh public safety oriented as it relates to these types of things. Uh the US particularly in the current administration is heavily leaned into innovation and winning the battle.

[00:22:38]And by the battle it in in in US terms that's around chips, that's around power and that's around models. And uh and I think you're unlikely to see at least in the current incarnation of the political environment in the United States. Much more regulation happening there. I think you'll continue to see unfettered unfettered innovation.

[00:22:59]>> But it's also great what kind of innovation is coming out of the US right now. I really love to seeing it. And we talked about it like the startup scene and innovation a little bit before the we start the recording. Maybe we can tackle a little bit about your background, like what drives you, how did you get into the startup space?

[00:23:19]>> Yeah. Uh yeah, when we were talking before we started taping, it was fun to hear your story and the number of startups you're working on. So, uh my first company I I started with my brother and uh and a good friend of ours, Mike Hilton. Uh we started that when I was 23. That was Concur. It was a travel and expense company that ended up being a global company. We did it for 20 plus years. And so it went from startup in my apartment to uh a public company to sale to SAP actually. Uh the the second company was another startup but a healthcare US-based healthcare company that was you know virtual medicine. So virtual primary care physicians, nurses, etc. Uh I love building stuff. Uh I I think I'm old enough now to kind of know what I love to do. I love building products. Uh the rest of company building is also fun for me, but the most part of the story, the best part of the story for me is building products.

[00:24:18]And so the smart sheet opportunities, we started at the outset. We talked about this Sasha. Uh I took it in in large part because I believed, oh, somebody needs to come rebuild this product for the new world. And what if you could do it with, you know, north of a billion dollars in revenues and hundreds of millions of dollars of profits and customers who you could experiment with?

[00:24:42]I'm here in New York right now and I'm meeting with two of the largest fintech companies in the world uh in the next 24 hours because they're so they're use our software and they want to they want to build on our AI platform and like that just is like you have a bunch of tools.

[00:24:57]So long story short, I love building products. I love startups. This is not a startup, but it's a kind of a startup moment, you know, and uh that's why I'm having fun.

[00:25:09]>> This is like kind of a general observation right now. I think there is has never been a better time to build a startup than right now.

[00:25:17]>> Are you is it correct?

[00:25:19]>> Yes. Yes. I mean, you got three going, man. Uh yes. First of all, the cost of writing code is as low as it's ever been, especially if you don't have a code base. Meaning, it's one thing to inherit 50 million lines of code or 10 million lines of code and then have to refactor it and rebuild it. Like your pace of innovation is a little different there than it is if you're starting from zero.

[00:25:41]So that's part one. Part two, the model's changed so dramatically every, you know, few months that your capacity to build things you never thought you could build. It's extraordinary. and access to data like MCP was a gamecher for the industry because all of a sudden data became available to every software company uh to the degree you're smart about how you did it like those factors in tandem you know the access to better models access to more power new chips are coming out now like the Google chip is going to drive cost of consumption way down which is going to drive the the power of the models way up yeah all in Sasha like this is just a great time if If you want to build a software company and you you we were talking about this before, if you really have the if you really have the desire, if you if you got it in you to build a startup, not everyone does, but if you got it in you to build a startup, now's the time to do it.

[00:26:38]>> How should you start? How should someone start who wants to say, "Yes, I'm all in. I want to build a a software right now." Where would you start?

[00:26:48]>> I think first you got to know something, man. I uh my last company was was funded by Andre and Horowitz. And Ben Horowitz said something to me when we uh when we uh actually did the first funding round, he said, "Great founders know a secret that nobody else knows. And when they know that secret, they go build their company around that secret." And so I think the best companies are founded around this problem you have that no one else has solved, that you've been trying to solve for yourself. like the what the company we founded uh Arty Semed is the CEO of a company called uh Scola that I'm the exec chair of uh we worked together he was the head of product management I was CEO of the company we worked together to try to find an AI answer to a problem we couldn't find it so when we sold accolade we said let's go build that company let's go build that let's go let's go answer that question and so I' I'd start with the question what burning problem do you know exists that hasn't been solved. Uh don't build a company just to build a company. Build a company to solve a problem and then uh and then do it the way that sort of that abides by your code.

[00:28:00]>> Yeah, we see also like a lot of one person companies right now because you especially in the beginning you don't need the workforce like how do you see the the workforce shifting then the next five years?

[00:28:14]>> Uh I might be a contrarian on this one Sasha. I think you can get away with one person and you know everyone's talking about the 10x engineer who can write you know you know 10x more code I think senior engineers can write a lot more code than they used to but once you start servicing customers and once you start delivering value and customers have requests and then you got to especially if you're in the enterprise software business you got to support those customers you're going to need engineers uh I I I do believe the employment environment will flatten out a little bit. But I think these this this universe of like all the, you know, all the jobs are going to go away is a little bit overblown. I mean, I know what's happening in my business. We have 3,000 employees. Uh we're probably not growing headcount as fast as we might have in the past, but I I don't see a universe where I'm going to cut the workforce by a half. Like, well, like we're seeing, you know, we saw Block do it. We saw other companies out there cut their work forces dramatically. I honestly believe that's less about AI and more about cost cutting for those businesses. Uh the the appetite for great engineers will continue to exist moving forward. And we're going to still need people to build things and to support customers. In the consumer space, that might be a little bit different and maybe you can you can get by with lighter uh engineering staff, but in the enterprise space, I don't see it.

[00:29:40]>> Yeah. There was also like a few funny press uh statements the the last few weeks where bigger companies are hiring again because the token costs exceeded the engineer costs.

[00:29:52][laughter] Yeah, like there it's it's true. Uh you know, you've you've all heard about token maxing and this AI slop. Uh there uh we will find over time that uh that the return on investment for AI as it starts to get measured wasn't quite what we thought it was in the last couple years. The next couple years it will start to turn in the right direction because business will be savvy enough to understand how to measure it. It's one of the things we're doing for our customers at Smart Sheet is is is helping them measure the AI projects that are happening inside their business by linking together their financial systems and their systems of work to get that to to actually do that measurement.

[00:30:35]I I think the reality is Sasha, we're going to look back on the last couple years as a couple years of experimentation. Everybody just went wild.

[00:30:43]>> Yeah.

[00:30:44]>> And you and I did too. Like I I I spend more time I I flew all night. I was on a redeye to get here to New York today.

[00:30:51]Uh, and I was screwing around with Claude working on uh working on different components of our product story. Uh, because it still blows my mind every day what I can do there. And so, >> yeah, >> the last two years are going to going to we're going to look back on it as years of experimentation. The next three or four years will be years of significant return.

[00:31:13]>> And that's also what the I think the companies are experiencing right now.

[00:31:16]they're experimenting and that that's kind of the issue as well. I mean especially in Germany you have a lot of midsized company the German middle stunt and many of them do not really know how to implement AI into their existing models.

[00:31:32]So what they are doing is they are just switching on um an LLM for their employees. So you now you can use open AI or anthropic or whatever and they are hoping for a fast return but this is not how it's working right.

[00:31:48]>> It's not how it works. I mean you you can improve individual productivity Sasha like we you and I can be smarter using uh an LLM. I I know for a fact that uh I was telling a friend the other day, I can read deep engineering papers that I could not read five years ago now by, you know, partnering with my eyes with the paper and with Claude to help me translate things into my own vernacular or my own terminology.

[00:32:17]Getting teams to be more productive, getting organizations to be more productive requires the capacity to leverage the LLM but with governance, with collaboration capabilities, and with context. And that to me is the Smart Sheet story. Like our job is to help companies understand the LLM can make you more productive individually.

[00:32:37]The LLM plus smart sheet can make you more productive as a company. And that's why we that our embrace of of MCP, our embrace of these models inside the software is all built around that idea.

[00:32:48]How do I take all of the governance and auditability of our software, all of the context of our software and deliver it to companies and say this is how you make AI productive?

[00:32:58]>> I mean it all starts with the data, right?

[00:33:01]>> I mean many of the companies first need to get their data right. It's just not every time but very often it's very chaotic working in different kind of of um of storage cloud.

[00:33:15]>> It's really really chaotic in these in these companies. They need to get the data right and then they can think about an in about an AI strategy and that's most of the time needs to happen top down.

[00:33:27]>> Yes.

[00:33:27]>> In my opinion.

[00:33:29]>> Yeah. Sasha, it's it's a moment for engineers to run companies or to have your point around data is so spoton, but most CEOs I don't think know that their data is a mess. Uh they hear the complaints like I you know it's hard to get access to the data or I don't have a single source of truth for this and they say it go fix that problem. But what really needs to happen and look you're certainly seeing it you know companies like data bricks and snowflake are exploding.

[00:33:58]>> Yeah. uh that most companies are figuring it out now, but it's maybe slower. People are are less apt to understand until you invest in a data bricks or a snowflake, uh you're going to have a hard time actually leveraging the full value or the power of AI.

[00:34:17]>> Yeah, that's that's true as well. So maybe a few words about you personally.

[00:34:21]Um you said you are also in a few boards like what other companies are you involved right now?

[00:34:27]uh a company called Scala that works for it really thinks about reinventing customer experience for contact centers.

[00:34:35]It's essentially uh how do we turn all of the data that powers a contact center into a proactive means of driving a better customer experience. Uh I co-founded that company with a guy named Marty Semedi who's the CEO. Uh they're cranking away. They just raised their seed round and uh from uh Madrona and Fuse in Seattle and they're uh uh they're you know on their first set of I think they've got about 10 customers and a couple you know and and growing. Uh and then a company called Basalt. It's a healthcare company. I was in the healthcare space for 10 years. Founder is a guy by the name of Ben Hackett. I'm on I'm on his board. I'm an investor in the company. uh and what they're doing is leveraging AI uh to make doctors more productive by uh by summarizing all of the different sources of data and delivering them to doctors so they can deliver better care. Uh all of it's really built either one of these solutions is built around how do I take AI and leverage the data that exists to make operations more effective.

[00:35:37]Yeah, I think especially pharma and and and far and medicine is like one of the most important cases for AI.

[00:35:45]>> Oh yeah. I mean, think about it right now, Sasha. like uh Mustafa Sullivan uh you know the deep mind founder wrote a book called the coming wave uh probably a year or two ago and he talked about >> the advancement of models protein folding you know sort of the advancement protein >> uh the advancement around gene altering or gene coding uh and uh robotics like these four waves are fundamentally going to change the entirety of the world, the universe that we live in. But two of those are directly healthcare related and they're very correlated to AI and it's moving so fast. I personally built me a whole um supplement stack for longevity together with my AI going through all my blood results and everything. And imagine that two years ago it would be would be impossible.

[00:36:43]>> 100%. I do the same thing. That's very funny you do that. Uh, I do the same thing uh for about for about two months.

[00:36:50]I just I recorded everything that I ate and then got a bunch of blood tests, put it in and then uh started building my own stack with my doc. But I but started building >> I was sending >> Sorry. Go ahead.

[00:37:05]>> I was sending my doctor a list of u of blood tests I need to do and I got a phone call. Why? Like where did you get that from?

[00:37:15]That exact same thing happened to me. I thought I was asking for stuff and she was like, "What do you I don't even know what you're talking about. I have to go do research >> to find out what you're talking about."

[00:37:25]>> Uh, yes, that's true. I hope she's not listening because I thought, "Oh, I wonder if I need a new doctor."

[00:37:30]>> Yeah. But I think like it's so powerful like the whole I I think the the biggest mistake you can do right now is dying within the next 20 years. I think when you can make it past these 20 years, the chance that you're extending your lifetime is very very high.

[00:37:48]>> It's very high. It's very high. I agree with you. You know, like the I think the advancements with AI and with the the the sort of the the the things I already mentioned uh will extend lifespans significantly.

[00:38:04]We'll cure a lot of cancers. Uh we're already making great progress on things like childhood disease, leukemia. Uh you're right. You're right. Try to stay alive 20 years.

[00:38:15]>> Yeah. [laughter] Working on that. So you tell what would you tell uh a founder today? I mean you I think you said said once speed beats data. But on the other end with smart sheet your thesis is data is beating speed. Like how does that fit together?

[00:38:33]Uh, I'm a CEO, man. Just like you. You're a founder. You get it? Like, uh, I I preach speed every day because the world is moving really really fast. And especially when your company gets to like 2,000 or 3,000 employees. Uh, you you start to get too many committees, you have too many people who want to meet about stuff, and you just have to preach like make decisions. Uh, the the point around speed beats data is really more uh decisions beat indecision every day. You're not you're there's almost nothing you can do to kill the company with one bad decision, but a whole bunch of indecision will kill a company. Now, on the point of data as it relates to what Smart Sheet does, look, projects are run in these moments between departments where we're not sure who gets to make the decision where uh we're unaware of the context of what's happening in that department or in this group. uh Smart Sheet has seen of you know we run 100 million 180 million automations a month for our customers like those that's 180 million workflows all of that knowledge of how what has worked which resource constraints are going to kill your project uh what dependency is going to delay things and cost you $100 million down the road those are things we've seen in our software and we should be able to deliver that to our customer agentically we ought to be able to deliver to their agents or the humans and that's exactly what we're building.

[00:40:00]>> You've hired from many companies. Um I think right now like what is the the traits you're looking for when you are hiring new employees >> right now?

[00:40:10]>> Yeah, right now.

[00:40:11]>> Right now, curiosity. Uh you got to be curious. I ask people every time I when I interview them, what are you reading?

[00:40:19]Uh and like what are you listening to?

[00:40:22]What are you reading? What's the last uh great book or article you read about the advancement of AI? Uh I want curiosity.

[00:40:32]Like this world is different than the world was two years ago. And if you're just trying to do what you did two years ago, uh you're going to kill us. And so I I don't want people who are comfortable. And then you got to I I I interview for relentlessness. I I like to meet people who have been through failure and then gutted through it and got it to the other side. And that's startup life. Uh I think startup founders or people who have lived in the startup world understand that a lot of days you wake up and the world is kicking your ass and you got to fight through it and get get to the other side. And so I look for relentlessness, diligence, like the capacity to work and I look for curiosity.

[00:41:13]>> So what would you say to someone who is studying right now or who is going to study soon like what topics should they focus on?

[00:41:24]Uh I still think engineering is a great field. One of the things I think the world requires is the capacity to continue to debate. Like you and I are having a conversation and if we were sitting over a beer, we'd have a chat and there would be things we disagreed on. And that disagreement is good for the universe because it it actually progresses pushes ideas forward. the framework about that debate is it is I think there's a there's a lot of sort of engineering framework about framing a hypothesis proving the hypothesis and actually delivering value. Uh how much how many times do you use an AI model right now Sasha and it tells you how smart you are. Uh, it tell Oh, that's a great question, Sasha. Thank you. That was so smart. Like, I can't believe >> the world's going to need the capacity to think and to disagree and like AI models are fundamentally revenuedriven, right? And so, they want you to keep coming back. They want you to like them.

[00:42:23]Uh, in business, we don't need people to like us. Uh, we need our customers to like us, but inside the business, we don't need people to like us. We need to have the right idea, and that right idea needs to progress. So you could study philosophy, you could study engineering, everybody's going to be an engineer now, right? Like everyone's going to be able to write code. If that's true, like study the capacity to build a framework, create a hypothesis, and actually question and value the other people's hypothesis. Uh that's what we need to do now. Machines are going to do a lot of the rest. We still have to do that.

[00:42:57]>> Are we losing that a little bit right now? the ability to talk to each other to discuss because I mean that's kind of what I'm seeing especially with the younger generation. I mean we are both also not that old but with with the younger generation with the social media going on a lot of um the communication is digital. It's not like face to face like it used to be. Are we losing the plot here a little bit?

[00:43:22]>> Yes, we are. Yes, that's the bottom line. I I don't want to be negative or I don't want to be pessimistic.

[00:43:29]uh but the capacity to actually communicate with each other. Look the political environment across the world is is polar polarizing and uh and confrontational versus uh philosophical and debate oriented and uh and that extends I but I think if you're building a company man like this is why I love building companies. If you're building a company you get to pick the culture you want inside. You get to decide, are people going to come to the office and actually show up and meet each other every day? Are you going to build a culture that's collaborative and that it welcomes new ideas and that welcomes that diversity? Uh that has debate and that doesn't just sort of uh that pick a lane because like by via some ideology. Uh we have a chance to impact how the world behaves by what we build.

[00:44:20]If someone looks at Smart Sheet one year from now, what would he see on the platform? What he would not see today?

[00:44:30]>> One, you would see uh that agents are driving the majority of our workflows.

[00:44:36]And so, uh everything that used to be done in uh in a hands-on keyboard configuring the software, our implementation would be will be deliverable. So you if you want to configure your own complex workflow and build an agent, you can do that in smart you'll be able to do that in Smart Sheet without having to talk to a professional services person from Smart Sheet.

[00:44:58]That'll be new. Uh if you want to instrument a brand new workflow and connect to 200 300 data sources, uh you'll be able to do that on your own without talking to anyone at Smart Sheet. That'll be new. Uh and uh finally we'll keep building underneath the hood capabilities meaning via and exposing them via our APIs and MTP uh for more power to be in your hands conversationally. You'll see a visualization engine as well so that you can you can speak to the software from your LLM and then see it the workflow visualized right in that LLM.

[00:45:35]And that's that's something you'll you'll see from us as well. So you keep on adding AI futures uh to to to the product.

[00:45:46]>> Yeah. I mean Sasha, the future of our company, the future of every company is agents are going to be using your software. And so you have to build for agents. Uh you have to build for humans, but you we've been building for humans for 20 years. Uh we know how to do that at Smart Sheet. uh building for agents and how agents are going to use your software and how much more effective they're going to be using your software and how much more they're going to demand of your software like that's the future and uh it may be a hybrid world for a while but the bottom line is that's the future.

[00:46:18]>> How do you see like the the current state of the economy? I mean there's a lot of debate if it's an is AI in bubble or is it not a bubble but obviously the market moves in cycles right but I don't see a bubble yet because there is still so much going on and so much things to improve I don't really see that argument >> yeah there's a lot of room here's what's happened up to now like I I don't I wouldn't call it a bubble so I agree with your point around a bubble.

[00:46:52]Here's what's happened up to now though.

[00:46:54]Uh huge swaths of the indust of of industry have dropped in value dramatically.

[00:47:01]That's let's call enterprise software one of those categories. Now that has to happen. There's a finite amount of capital in the world and if it's all rushing to the AI companies who are now worth hundreds of billions of dollars, then that capital has to come from somewhere. uh and obviously semiconductor companies are way up like these companies that are sort of powering this new wave are way up. Now the market sort of writ large just rode off one whole sector of software and what will happen is some of those companies will come back up to the top.

[00:47:32]Why? Because they're growing because they deliver value in the world of AI and they're very relevant and they will continue to be very relevant. some of those companies will go to zero and so instead of all of them going down some will go up and some will go all the way down that will shake itself out over the course of the next two three five years and uh when it does there'll be a a return to some normaly right now there's the market just moves in waves and the wave is invest in AI invest in power invest in semiconductors etc and uh and everything else has to suffer >> what's your prediction for the economy I think Musk said 10x the economy till, don't get me wrong, I think he said till the next 20 years or something. Didn't get the don't remember the number.

[00:48:20]>> I don't remember the number either, but uh that sounds very much like Elon Musk.

[00:48:24]I think the uh uh I'm optimistic about the economic growth rate. Uh I am uh I don't you know I'm not an economist man. I don't know is it is the is the world GDP going to grow by 2% or 1% or 3%. I I don't know.

[00:48:39]Uh I do think that economic growth will continue especially as we start to see like AI hasn't been harnessed in a way that delivers economic value for enterprises, right? Like that's a giant unlock of the economy. When it really happens organizationally and corporatewide, which is something we're working really hard on our sheet, I think that's going to deliver enormous amount of value from a GDP perspective.

[00:49:05]Now this the follow on question though that you've already asked Sasha is like what's that do to the unemployment rate?

[00:49:09]Does the unemployment rate go up or does it come down? Uh I I think it comes it goes up uh but I don't think it goes up quite as much as everyone's thinking but you know like I said >> I'm a software companies uh entrepreneur not a econ economist but that's my guess.

[00:49:30]So you with maj you are serving a lot of different kind of industries. We talked about that before. What do you think like what industry will see the the biggest growth?

[00:49:43]>> Uh I think you're going to see healthcare companies grow like crazy. Uh life sciences, pharma, biotech, like those companies are going to grow like crazy. Like there's just so much room to the point of our our conversation earlier, there's so much room for innovation. we we do really well in that sector and we continue to lean into it.

[00:50:03]Uh I think some technology companies are going to win big. Uh those companies that really lean into the AI story and lean into their data and the value the the the the value or the mode of that data. Uh I think really interestingly service companies Sasha like if you think about service companies where the the margins of those businesses have been 30 you know 20 to 30%. The gross margins of that business in the 40s those businesses you're going to see gross margins go up to you know into the 70s probably you know it's certainly into the 60s maybe into the 70s. uh that will be extraordinary in terms of return to their shareholders. And so those are those are in my mind three of the big winners that will happen in the economy soon.

[00:50:53]I had an interesting analogy um from another guest a few episodes back. He said the application layer will always win and he said think about the time refrigerators get get uh out there. It was not like the people who built them who get rich. It was the people or the companies who are using him like Coca-Cola. It was the perfect application for the refrigerator.

[00:51:18]>> Yes. Uh I I think I think that's correct. I do I I think uh I would the the addendum to that point is the application layer will continue to be relevant if the data that's delivered by that application layer is unique and differentiated and accessible from a contextual perspective in the right moments. If those things aren't true, the application layer will become less less important. And so I understand the point of your previous guest building applications is just getting easier, man. It's just easier to build stuff. And if that's true, if so, if the cost of writing code is low, then the output of that code and the customers you've acquired and the data that those customers create is your mo that is your value. And you know, every company has some slide, every big software company has a slide. So, they got 20 pabytes of data and I got 20 years of the history, all that. Uh the the next question is like okay well what do you know about the universe that nobody else knows because if your data is largely generic and meaningless uh you got a problem.

[00:52:30]>> Yeah definitely. Is there anything we need to talk about what we haven't talked about yet?

[00:52:37][sighs] >> Uh I don't know man. I've had this has been a fun conversation. It's gone by really quickly. I think uh I think we've covered a lot. I'm uh we're really excited about uh about your audience.

[00:52:49]Like the market in Germany is a place we really really want to grow. I I intend to spend a bunch of time there. Uh I'm I'm uh I'm quite a big fan. Actually, I'm going to be there for October Fest this year. So, uh I'm >> Let me know. I'm not far away from Munich.

[00:53:03]>> Yeah. Okay. Well, I will let you know for sure. Actually, I will let you know for sure. I think it's we're going to be there that first week. We're having we're going to have a little uh corporate tent there. So, uh we should we should have a pint together for sure.

[00:53:14]>> Yeah. Let me ask one more question because this is a question I ask every guest. Um, what's your wildest prediction for the next 20 years?

[00:53:24]And I know that's a tough one because with all the stuff going on right now and how fast things are moving, it's really hard to make a good guess. But what's your wildest prediction?

[00:53:35]>> Uh, two things. I think we'll start to see tangible improvements in lifespan which we talked about earlier and like keep in mind lifespan in the United States is going down right now. Uh it's it's gone down and that's that's a different problem but I think you'll you'll see the lifespan start to increase again. Uh number one and number two uh I think economic growth will be all-time high 20 years from now. like we we'll be growing faster than we are right now which is uh which will be hopefully extraordinary for the world around us.

[00:54:14]>> It's exciting times. Data centers in space, huh?

[00:54:17]>> Yes. Data centers in space, man. Power from Yes. data centers in space, new chips, power from everywhere. It's going to be amazing. [laughter] But I I I do I I am optimistic just like you are. Like you you don't start companies uh and build stuff if you're not optimistic in some way, shape, or form. It's it's it's inherent in the actual premise of starting something that you believe you can finish it and that's going to take time. So, I'm optimistic like you are, man. But, uh, I appreciate you having me on today.

[00:54:46]>> Thank you very much. I think that's it for now. Great closing.

[00:54:50]>> Thank you.