Why We’re Only Using 1% of AI | Glean CEO Arvind Jain

The technology stack is evolving at a pace like one that we've never seen before. My mindset by default is that if you build something last year, that it's got to be obsolete. There has to be a new way to do that thing better today. And if not, then it's just lack of imagination. Startups in general are a lot of hard work. There's a lot of fun. You have people who are really motivated, who are aligned on the same mission. You really love that sort of feeling of we are in this race and we're going to build a great product and we're going to win, we're going to bring customers. being in the AI industry doesn't make it easy because this is the most brutally competitive industry that one could be in. Like there is so much change, so much competition that you don't feel like, you know, you can get off the treadmill to drink water. That's the feeling of being at the company is. But I think like, you know, it all works because we know that there's a great destination. Welcome to Grit. I'm Juven, partner at Kleiner Perkins, a show where we go beyond the highlight reel and explore the personal and professional challenges of building history-making companies. Today on the show, we have Arvind Jain, founder and CEO of Glean, the enterprise assistant and workplace search platform. Glean was actually started in the Kleiner Perkins basement. This is Arvind's second time on the show. It is incredible to see the progress that this business has made in just a few short years. A fascinating business to study and watch. Enjoy the episode. one of the things that has transpired over the last, whatever, six months or whatever, is I'm incubating a company here that I'm the CEO and one of the co-founders of. Oh, cool. Congrats. And thanks. And then we're growing. We're growing. And so I pinged Michael and I was like, hey, do you have space for eight or 10 people in the glean office? He's like, let me get back to And then he was like, no, no space. And I was like, damn it. We have it. I was like, no. Well, the office looks pretty empty to me. Okay. Yeah, we should. It's okay. Are you still looking? What's that? Yeah. Okay, I'll tell him. Are you going to take the downstairs? The first floor. We don't have plans, but you should come. That'll be cool to hang out. Wouldn't it be fun? Yeah, totally. I will tell him. Is there space? My feeling is that there is space. I don't think it's fully occupied. My feeling is that he may have said no because of little people. Okay. It's like, you know, once you have little people, like everything changes. Yeah. You're a big company now. Yeah. So that's probably, my guess is that that is the, that's probably what must have happened there. Okay. Well, if you think there is, I didn't want to bug you with it, but it would be fun. hey and maybe you know get some new startup energy into into gliel yeah yeah it's nice i like it yeah and i think like it'll be yeah it'll like make things easy for you like you know we have all the all the regular stuff like lunch and everything taken care of you don't have to manage all this like you know like you can even like you know get them to just put you into the day-to-day system for all the other employee needs, right? So I will tell him, like I don't know whether the legal will, or like, so like I guess they will tell whether they have space or not. Yeah. And whether legal will allow or do they need to sort of like create a wall or something like they market it. Yeah, yeah. Okay. But what's the company about? To make a long story short, pricing models have gone crazy. So for example, you all are thinking about doing a consumption-based SKU, okay? And even before you did consumption, you were doing all of these complex deals, early renewals, expansions, multi-year, more SKUs. The underlying software that these, like your team probably came back to you and is like, Arvin, it's actually really hard to do a consumption-based SKU. and a lot of the reason why is because the software that these things were built on is not designed for a world where pricing models are so complex like metanome or OR exactly so that's on the billing side but even on the quoting side like for example your AEs it's like very difficult so we want to do exactly so we want to do the full stack we're going to start on the configure price quote which is like basically natural language way for an AE to just say this is exactly what I want natural language way for an administrator to say, hey, we introduced a new product, show me exactly what's going to happen on the user admission testing. And then a kind of pager duty-like approval flow that has context relevant things for Michael to see something around COGS, CAC, LTV. Another thing for Deals Desk or RevOps to see, because that doesn't really exist in software today, which is insane. And obviously I come from this world and I've been dealing with forever. So that's what we're going to start with that specific land. And then we're going to go build out billing, contract lifecycle management. Very cool stuff. Congrats, man. That's good. Maybe I, so I'm glad to be doing this. I, maybe just selfishly, you're so big time now that I, I missed you. And I was like, well, the only way I'm going to be able to hang out with Arvid these days is if I, if I come up with a good excuse. So anyway, it's just good to see you. Yeah, likewise. No, I'm actually, I think this is, I was here last week too in the office. Cool. Yeah, I'd love to, you know, I'd love to hang with you. Like it's fun. It's fun to spend time. The San Francisco office. Yeah. We, I've been hearing about that since day one. Yeah. Day one, people, you've hired people from San Francisco. Yeah. You finally, what, seven years later? Yeah, six years later. You relented? Yeah. Why did you, did it break your heart? No, it didn't break my heart. It's a tough decision. And actually, like it was not so much that I was in the way of it. The main reason for us not to have an office in San Francisco for a long time was because we wanted the team to be together. And it was not like so much like, yes, I of course like the team to be together too. but our R&D team, our technical leaders, managers, even the ones who are living in San Francisco didn't like the idea of that, like people are going to be in two different offices. And like the one that you start at, like in Palo Alto, it's hard to get rid of that one. That's HQ. That's HQ and has to be because all these people who came to the company, they made the decision to join Glean. Part of it was that I'm going to be working from that location. So that has to stay. So then it was always a choice of like, well, you know, do you do one more location? And I was actually very motivated. Like the thing that I used to hate the most was people from San Francisco wasting three hours every day commuting. Like that's, that's like, you know, I want all of that time to go developing systems, right? So, so I was, so that was the motivating factor for me. And ultimately I think this year I just felt that, you know, we're going to actually be going even more to the office. Now we have four days a week instead of three. And I just didn't see us, you know, me asking our team to make that track every single day. So that was the motivation. That's why we did it. How big is the company now? How many people? We crossed a thousand people. No way. Yes. I was also surprised, you know, it was not a, just to be clear, it was not a feeling of, you know, celebration. It was a feeling of panic. When I saw that, you know, we're past a thousand because that's a big number. That's just so big. But why'd you panic? What about that number? Well, I think the... Like if the organization's bloated or can you do more with less? Are you turning into a big company? Well, I mean, like, look, you hear a single person, you know, billion dollar revenue company as like the new North Star for people. and so like everybody's trying to be really lean in terms of building their companies and we got to a thousand and I think the, I do feel, I do feel that we have, we could be doing a lot more, you know, with our team. It's not like, I'm not pointing out to individuals in our company not putting in their best. I mean, our people work hard. Yeah. But when you get to that scale, to get more things done, you have to work much harder on alignment, prioritization, having clarity of what the business should be doing. And all of that used to be easy. You could just walk in the room and say, we're doing this. And everybody heard it. And now with a thousand people across like a hundred different cities with our field team, it's just hard to keep everybody aligned and be on the same page. So when that happened, when we got to a thousand, that was the first emotion that I had that, oh, we got to be like really, really intentional about like how we organize our team now. Because organization was not, it was not a thing on my mind. Like it was always product, technology, winning a customer, not like how we're going to organize the company. But now that's super important. And actually we still haven't done a great job. So we have to work on it. Do you enjoy that part? Isn't this kind of in some ways every founder's dream and then also every founder's nightmare, which is like you always think maybe you'll get to a size and scale where you are solving enough customer problems that you hire a thousand people, but then you get to a thousand people and you're like, oh my gosh, like, what have I created? Like, this is a battleship. I can't, you can't even move your own company. Yeah, yeah. Like, and by the way, the things that got you, like usually the skills of a founder. Yeah. Like, I mean, you and I know each other pretty well. Yeah. You are not necessarily the person that's really keen to go put in a bunch of process everywhere, right? Like, that's not what you enjoy doing. Yeah, yeah. Actually, it's interesting. The... actually get more annoyed with lack of process. Because lack of process makes us do oftentimes way more work. I have to answer the same question a hundred times because we didn't write it up and made it available to everyone so that people could do that work without having to check with me. So process is actually an important part of, I mean, like you have to, you can't be like, like, I don't like, you know, like there's no such thing as a, like, I'm a cool guy. I'm a not processed person. And, but I can, you know, wait and all of that. Like, you know, that's not enough. Like, you know, to build a company, you got to really, a big part of being a leader is to be able to organize, is to be able to, you know, set processes, you know, set a way of working inside your company. But then your question was different, which is, do I enjoy it? And the answer to that is no. I know that we need it, but it's hard for me to really excel at that, to lead our team to actually be able to do these things. So that's true. that's true. Like, you know, I think I, I like, as, as the company grows, like, you know, you know, one has to grow with it, um, and force, you know, force like myself, for example, to really learn and adapt. Um, and that's what I'm working hard on trying to do right now. Like, you know, with, with, I guess with some success, but mostly failures, but I think you, you learn from them. And today, like every startup, every company that I talk to is like, wants to be Glean, basically. Like now it's like Glean for X, which has got to be both amazing, but also insane. Like even for me, I'm like, what? And my general reaction is like, having seen it pretty close with you. Yeah. It's like, are you sure? Like, do you know how hard it has been? to get to this point. Like, uh, I don't think like, even now they're like, wow, must be unbelievable being a clean. I'm like, well, the company is doing well, but like the lived experience at clean has always been that it's not enough, you know? And, and you do that for years. Yeah. It's exhausting. Yeah. And I just wonder like, you know, like now a thousand people, people probably come and think I'm going to go work at the, you know, that really cool, hot AI company. And I think there is like this dissonance between the reality of what it means to work at a company like Glean, which is like, it's exhausting. Do you feel that? I mean, it is. I mean, like, you know, startups in general are a lot of hard work. There's a lot of fun. You know, you have people who are really motivated, who are aligned on the same mission. And as an engineer, for example, as a developer, you really love that sort of feeling of like, just we are in this race and we're going to build a great product and we're going to bring customers. So there's something exciting about that journey, but it's also like it's like that. It's like being an athlete is not easy. You have to work very, very hard and you do get exhausted. Like people do get exhausted, like after four years, five years, the grind is not going to stop. Right. And, and the, so that's, that's, that's, that's actually, um, like it's been very true for us. Like, you know, internally, like is everybody's existence, like, are we happy? Are we um are we like always feeling energetic like you know it hard like you have to sort of really push yourself to actually feel that way And being in the AI industry doesn make it easy because this is the most brutally competitive industry that, you know, one could be in. Like there is so much change, so much competition that you don't feel like you can get off the treadmill to drink water. That's the feeling of being at the company. But I think it all works because we know that there's a great destination and we're making a big impact on that journey. And so that motivates us. That keeps us all together. and the last thing that I would say because of this very question that you asked me I've been always telling our team every week or every other week I actually tell them that this is going to be a tough, long journey in some ways that is no better tomorrow like it's going to keep feeling like this like that's what it means to be at a startup is, you know, be in the grind. And so you have to figure out something. You have to figure out something that makes today fun for you, not the tomorrow that you're hoping to arrive at. At this point, that grind has been going on for what, seven years for you? Yeah, six and a half. Six and a half years? What do you get to do? You know, like, meaning you can't leave. You can't just like go to another company. Yeah. You can't just like stop. getting better. Like you can't, you can't really not learn process and abide by it. Like you have to do these things or like the company goes down and like, well, you know, what are you going to like let, you can't let that happen. Right. Yeah. So I don't know, like rubric went public. You started rubric, like you've made plenty of money. Yeah. Like how do every day you wake up and you're like, I don't know, you just, you gotta like get ready to go do it again and again and again for six and a half years and there's no like end in sight. Yeah. I don't know. Like, does that treadmill get exhausting for you? It does. And I think this is why I was saying that you have to learn to make your day fun today. You know, I'm not waiting for that tomorrow where there's going to be a great outcome, you know, with clean and will succeed and suddenly will stabilize into a business that just that just works and it's easy for everyone actually this is like you know never happens to any business like it doesn't matter like you know how how big you are you can go public the complexity of a business and the greater the responsibility you take at a company is going to be exhausting it's going to be it never stop never stopping sort of um you know you know series of work and stress and all of that so so like you have to you have to like just figure it out like for yourself like you know why are you doing it what what's important for you what is it what what else could you be doing in that and so for me the i know that if i don't um if i don't have a problem to solve that will be a miserable existence for me i don't have hobbies you know to to go travel the world or like i would be really lost if i didn't have something to build i have to have to have something to build i think so i think i think that that's how it is i mean And for me, I feel that is sort of my, fundamentally, I need work. Like every moment of my, when I'm awake, I need to be doing something. But for me, it is fun. It is fun because one, like, you know, I'm mission oriented. Like I like, you know, what we're building. I know we're adding value. I know we're changing, you know, how people work and making them more productive and making their work more fun. um, uh, achieving things for them. I also really enjoy just spending time with my team. So sometimes like, you know, and I will maybe not go into the details of it, but some conversations totally sap energy out of me and some fully energize me. And so I have to sort of make sure that my day is planned so that I'm, I'm getting a little bit of both. Like I need to get both the dosage, the dosage for both of those. Yeah. Yeah. Yeah. Do you, do you feel like you don't have a bunch of co-founders that are helping you right now? Like you have an executive team and companies started with, with folks, but like it's your company, there's no way around it. Does that burden weigh heavy on you? Like this company is whatever worth, last valuation is like 8 billion or whatever. and it doesn't seem like it's slowing down anytime soon. Glean continues to accelerate growth and you're getting attacked from all sides. OpenAI wants to be Glean. Yeah. Flexity wants to be Glean. Everybody like wants to try and build a product like Glean. Yeah. And you have some really like hard decisions and I've had for a long time. Yeah. there is no precedent for being Glean I'm not sure who you even if you had other people in the company I'm not really sure you can go to them with those decisions there's no right answer does that weigh on you or not really I mean like it is it is hard to to run a company I don't know like how much of that is like like running Glean is extraordinarily hard compared to others, it's probably the same. I would say like every company will have some different challenges. Like, you know, for us, the, like, while, you know, we have a lot of competition that is coming up, but also we had a phase where we had no competition for many, many years. You know, in those days, you know, our problem was that we had to really evangelize and convince people that like, you know, that a product like Glean is going to actually help them, help their employees. Now we have no convincing that is needed. Not just us, but 15 other large software companies are educating the rest of the market that you have to invest in a tool like Lean. So we have a lot of demand, but we have to compete. And we have to convince customers now of a different thing, which is that we're better. And I think we are able to do that because we have more experience than anybody else in this product, in this domain, we've been here, we've been at this problem the longest. So I think like in some ways, what I'm saying is that, like, you know, opportunities, like the opportunity keeps increasing for us, the challenges keep increasing for us too. And it is tough and it is hard. And, but I have a great team. We have people like who've been there from the company from day one, like, you know, my co-founders, my core part of my founding team, you know, our executives. And we are all here to make success. And I think we will because we're focused. We are this area of search and enterprise AI platform. We are more focused than anybody else on that. And so we feel pretty good about our chances to keep succeeding. How insane is it that like the ground shifts underneath you? You're right, like for years. Yeah. We were the only game in town. Yeah. And there was a goal. Every day you chip away at that goal. Yeah. Their LLMs didn't exist in the current instantiation at that point. Yeah. It was connecting your applications and then giving you visibility across everything in your enterprise. Yeah. Easy to explain, easy to understand. Yeah. The applications were all relatively consistent, you know? Yeah. Employees, you turn it on, they love it. Yeah. Nobody else was trying to do what we were doing. Yeah. Now, OpenAI is trying to do it. And new models are coming up. People are now trying to build agentic-related things on Glean. You're now competing with 100 other vertical agents. That rate of change has to be insane. You're really one of the premier AI companies. and the tax of that is that your rate of change is like higher than anything ever. Yeah. Like there's a lot of pressure on all the teams, you know, in the company. Think about R&D organization. There's always this sort of challenge, you know, when you think about building a new feature, a new component that, well, should we do it or not? Would the LLMs just get smart enough to do these things on their own? And so we do live in this world right now with AI, where the technology is moving so fast. And it's sort of like you can't even think of that, hey, like you're building a technology mode, you're building things in a certain way, which is going to give you a lot of strength over time because things change so fast, so quickly. In fact, I think whatever you think of mode is probably liability in some ways because you have to evolve, you have to move fast, change your product, change your technology to adapt to the new sort of foundation, as you mentioned. So it is hard, but I think we have a good team. Our engineers are capable of working in this shifting foundation environment. And so far, so good. Like, you know, I think people do love the clean product. You know, they constantly I was just, you know, like before this, I was in a call with the largest PE firm, one of the largest PE firms in the world. And they made the same comment that like, you know, like this is search and an AI platform is a priority for us. We went with the large players and it didn't work for us. and you are already in many of our portfolio companies and they love it. And so we're excited to work with you. And so there are those moments that also happen that then sort of bring back that motivation to everyone that let's keep winning. Your comment on the moat, the minute you think you have a moat, you're probably in trouble. Yeah. Is that because you just fundamentally aren't sure that that's possible when the LLMs are getting better so much faster? Yeah, that's part of it. the technology stack is evolving at a pace like one that we've never seen before. So like my mindset by default is that if you build something last year, that it's got to be obsolete. There has to be a new way to do that thing better today. And if not, then it's just lack of imagination. And so we, like one of the things in which I keep telling people in our R&D team is that reward building new stuff as much at the same level as throwing some stuff out of your code base. Because if you're not throwing things out, it means you are slowly and slowly converting into a legacy software stack. so the that's that's a that's a that's an important thing and that's why this concept of more to have to think differently it's not the code that you wrote what is it well i mean i think sometimes you know it is um like how quickly you can remove and replace code well one of them is agility yes so the a new currency now and well i guess it's always been but it's more important than ever before is how fast you can adapt, change, change your product, change your code. That gives you the leverage to take advantage of the latest technology that's being built out there. Then your mode is also your relationships, your customer relationships. We work with the largest enterprises in the world. and we are a core partner to them in their AI transformation journey. And we're very clear to them that we're not actually coming with, shipping great technology to you. That's not going to be enough for you to drive that full impact with AI inside your enterprise. You actually need a partner. You need to work with somebody and we're going to be that partner and be on the journey with you for a long term. We'll bring the AI expertise. We'll bring the shared learnings from our other enterprise customers, but we will spend the time and we'll keep working with you. And so it's that working relationship that is also more to our customers, love how we actually come to them, how we engage with them and how we build shared roadmaps to drive value in their businesses. Do you think that this idea of the code that you wrote a year ago is probably obsolete now. Let's take, let's extend that to like most SaaS companies today. Yeah. How do you apply that to a, like, are you in the camp that basically most of these traditional legacy players, anybody that was born like 10 plus years ago is in big time trouble because they have a big customer base that for them to, they have an innovator's dilemma. Like they probably have to rewrite their stack because it's not as good as it was. They could do it way better. But in order to do that, like that very cannibalizing to their existing business Yeah Is that how you think about it Well I think so two things You know one is people have talked about that there going to be no application interfaces There's a database behind the scenes with all their data. And then there are these modern AI experiences which are conversational and all work happens like that. And I don't subscribe to that. I feel like, you know, there's going to be a lot of different product interfaces, different products that are going to continue. Enterprises will have probably more products rather than fewer even in the future. So I don't think if you are a SaaS product, that suddenly you become worthless, that you're not important. But that doesn't mean that you don't have a challenge. If you're a SaaS company, you cannot stay static. You have to really go and innovate, update your product capabilities, build a lot of native AI features within it, which sort of like meets the bar of what's the new expectation that users are going to be having. But I think I do feel like they will continue to have amazing vertical solutions to different business problems that SaaS companies are going to be providing. It's just that they have to innovate. They have to take advantage of all the great capabilities of AI in their new product experiences. And I don't know if it is actually necessarily cannibalizing. I think you can add sort of AI can be additive to you in terms of what to deliver to your customers today. Do you think that seeing what you're seeing and being probably one of the technologists that I most admire, do you think the music is going to stop anytime soon? Do you think that we're going to keep seeing this thing? It's getting hot. Things are getting hot right now in the valley. Do you think it's going to keep getting hot from what you see in the technology? Yeah. Well, I think on the talk of AI bubble or not, or valuations, I'm no expert, so I can't actually... Yeah, not even valuations. The models keep getting better at the same rate. Yeah. I think models, yeah, I think technology will keep getting better. In fact, I think the... It's sort of... amazing that like that same core technology, like we continue to scale it, make it, you know, make it work on more and more data and with larger and larger systems. And there hasn't been a fundamental sort of new technology that has come as an alternate way to sort of build these models. And so, and that's also, that's also going to come like, you know, that's the human innovation, you know, spirit, like, you know, we will see a lot. So I have no doubts that the, that AI capabilities are just going to go increase more and more over the next few years. But even more important is this concept of how much are we even leveraging what AI can do today? And I would say that we've not even used 1% of current capabilities of these models, not even 1%. Today, the way the models are is that you do need to do a good amount of work on top of them to truly build products or solve problems that generate business value. so so I think like I see like you know I can imagine a world like imagine that like there's no innovation in AI models let's say they stay static where they are today even then you can you can expect that in the next five years you will have massive growth in AI powered products across all industry verticals I will get from that one percent of the current capabilities that we're using, we'll get from one to 10 to 20%. So that's, we have to remember that. Like, I think we sort of sometimes get carried away by will the model core base model improvements, you know, will they slow down or not? Like it's not relevant in many ways to most business problems today. Do you have a frame of how you play? Like in some cases, these model providers are both your best friend and your competitors. Yeah. Right? Like, and by the way, like, I preferred the world when there was no competitors, not when you're competing against these people. These are not, this is not Joe Schmo competitor. Oh yeah, we are the world's best companies coming in this. Trying to do Glean. Yeah, it's one of the things that they want to do. I mean, obviously all of those companies are much bigger than us. They have great capabilities and this has become one of their important projects too. So the... It's a delicate dance, isn't it? Yeah, yeah. I would say two things. You know, we are close partners with all these model companies. We collaborate on a technical front. Like we'll tell them, for example, what are the areas where the models don't do a good job? And so we have really good collaboration that way. We are, of course, they're customers too. Like we drive a lot of usage and a lot of revenue to them. this concept of co-optition is sort of so pervasive now because I think everybody, every company today is increasing the surface area of their products. Everybody's trying to do more because AI is actually allowing them to do more things. So overlaps are also increasing. And, but I feel like there's a, it's a bit of a temporary phenomenon and I think everybody will find their swim lanes again. And the way our company is built, for example, I feel like we should be fully complementary to all the model providers. We solve different types of problems. And there's a time when we were training models too. We still train some small ones. And my feeling, my guess is that we'll stop doing some of that work. The model companies will start doing some of the work. Why? Because ultimately you realize that you cannot do all the things. You have to focus, you have to go deep to compete. You have to get really, really good at certain things. And if you're trying to be everything to everyone, then you just cannot compete with somebody who's focused on a smaller problem problem and going deep into that. It must be an interesting moment in time for you seeing what, let's just say, OpenAI is doing. They're so ambitious as a company. Now, that might not pay off to your point. They might narrow their ambition again. But I wonder if it makes you ask, are we doing enough if they're so ambitious? It's a weird thing because you're like, wait, but we got to keep doing our thing the best. But boy, it's impressive that they're so ambitious. I didn't even mean to say that they will reduce their ambitions or narrow it down, but I think like they will probably like over time, like any company, you know, you make a lot of bets and then you start to consolidate your bets and you start to go deeper into the areas that are the most important to you. And so that's sort of, that was a comment that I was making. So I think that will happen. In terms of whether we are doing enough, I feel like, you know, internally, so we listen to our employees and our employees actually have the reverse complaint, which is that we're also trying to do too many things. Like we're a horizontal AI platform. We connect with all the different systems. We can come and build agents for your HR teams, for your IT teams, sales, customer success. And so the problem is, I think, the reverse for us, which is like, are we trying to bite too much? And we're sorting it out. Like, you know, one of our strategies is to be that horizontal player and have like deep partnerships with all other vertical product companies. Yeah. And like how do we sort of, for example, like our goal is to make sure that, you know, if you're using a CRM product, if you're using a, like, you know, like a, you know, some engineering system, within those systems, you know, we want to actually add some value. Like we want to actually build a platform that provides, you know, the deepest, the broadest enterprise context to AI within those, each individual systems. So that is sort of like, you know, what we are thinking about in terms of our strategy. And so that we don't have to solve all the problems. We don't have to actually build the best products for every function, for every department, for every vertical. But instead, you know, we become this platform that powers, you know, those experiences. So we do a little bit of work in that. We provide value, but we keep our focus on being a platform. Has the way you've thought about talent with these, let's just take engineering talent, with Cursor and others, has it reframed the way that you evaluate and think about talent at Glean? So one thing that we've been trying to do is we are testing these days for AI fluency for all roles. And the goal for AI fluency is not that are you an AI expert? We're not expecting anyone to be that. We're trying to see like, are you inquisitive? Are you like, you know, this revolution is happening in front of us and how much interest have you taken into it? How much curiosity it generated in your mind? So we want to actually see what have you done with AI? How have you used it? And so that's part of like a change in our interview process is to look for that. Like, is somebody going to be AI forward or not? Is somebody going to be, is somebody going to have that mindset of like, I want to do things differently now, not the same way that I've done before. So that's the one change that we have, that I'm trying to actually bring about, like in every team. But then at the core, I think like how we recruit, I think it remains the same. You always are looking for people who first... I mean, we look for all the things that everybody else looks for, which is you're looking for somebody who works hard, somebody who's talented, somebody who's a good team worker and all of that. But I think we always tested for one more skill, which was important for us, which was how passionate they were for the product that we were building. It was very important for us to see how much... And that's a test for us, whether they want to be aligned with the mission of the company and stay with us as we go through the ups and downs. Don't name names, but just in general, I'm curious, senior engineers that are at Glean, that previously before LLMs were obviously the best engineers. Yeah. Let's just say they're 20 years of experience. have you noticed any changes in how those people are valued in the organization relative to the person that's been basically grew up on Claude and Cursor? Yeah. Like, how do you think about that now? Because it used to be pretty clear cut. I'm curious how that's changed for you. Yeah. AI adoption is, so our, my co-founders, Tony and Vish, they are you know for them AI usage is a yardstick like you know their own assessment now is that if somebody is not using these tools it shows a mindset that is not open and so they really value everybody taking more and more advantage of these AI tools the patterns that we see if you look at the younger folks people just coming in and their first job is at Glean they're going to be a power user of AI because like you know I guess luckily they don't really know how to do their work yet so they don't know what is the traditional way of doing it so they sort of rely on AI to sort of help them So that is actually great. The young generation, they're going to be automatically become AI first. And then in our experience team, we saw this. We saw the power curve initially, and now we see two user populations, one that is very actively using it, one that is not so actively using it. And we're trying to understand. And we're not seeing that much of a difference in productivity between those two sets. Some really good engineers, they haven't adopted AI that much. But they're still productive because remember that AI has become helpful in terms of writing new lines of code. But a lot of work that the most senior engineers do is not writing new lines of code. It's about debugging, diagnosing, troubleshooting issues. It's about thinking and designing components. And so even like you know not using these code gen tools that like they still able to be our best people Arvind earlier you mentioned about like process and writing how writing is important to you. You and I have been in many of interviews where the candidate will ask me, what's it like with Arvind? How is he, how does he work? Yeah. And I will always tell them that before you talk to him, put something down in writing so that he can collect his thoughts and understand and observe how you think. And how you talk is not enough for him to see how you think. He actually needs to see it written, both because it helps him understand your process, your thought process better. And it's a way for him to digest his own thoughts and come back with Yeah. Clear questions. Yeah. Is that a fair characterization? That's actually, I'm surprised that you know that deeply, like, you know what I need in some sense. And we've been at this for six years together. I guess. Yeah, that's true. So I think there are two parts to it. One is my own limitations. I need to be able to read. I mean, I likely have attention deficit disorder, whatever it is, but I cannot process a lot of information if I have to listen. I have to read. So that's sort of like one reason why I prefer things that way, like for people to write for me. But it's also important for everyone. I think writing is the best way to really activate deep thinking. Like for everyone, I don't think it's a thing for me. And I learned this, like I've some of the best people in the industry, even when I was at Google, I would hear people talk about it. That like the to truly communicate effectively to actually make strategic decisions. Unless if you're going to actually take a pen and paper and write down your thoughts, you're going to be missing things. You're going to be, you know, all over the place. You won't have a coherent way of communicating. So I think this is a skill that everybody needs to develop. And this is where you got to really think about what is the role of AI? Is AI going to sort of, like given that now AI writes most of the things, like is that skill gonna go away or not? I have some thoughts, but anyway, but you're right. Like, for me, that is important to see things in a written form. What are your thoughts there? For AI? Yeah. I think AI can give you a good start and it sort of gets you past that initial hump, whether you call it the writer's block or whether you call it not having enough energy or time to put thoughts together. So AI can be a very helpful companion in terms of, you can say that like, look, like on any, like I do this actually myself now. I will sort of give some rough, unorganized thoughts to AI and say that, hey, put that in some sort of a structure for me and I get some artifact and then I start to work on it and now start to go deep into sort of providing my thought to it. So I feel like, you know, there's a, there's a, there is that world where people will learn how to use this, use this technology the right way, not, not a way which sort of creates this, what, what's called work slop, you know, where you produce like copious amounts of information through AI, but then you're not the reader, you know, just letting some, you know, pushing that out to other people to go and read, you know, that, that is sort of one, one, like, you know, bad thing that's happening with AI these days, which we need to stop. Um, one of the other observations that I have made about your style is, um, you have a real allergy, in my opinion, tell me if you think I'm wrong, to anybody that has accomplished anything and leans on that accomplishment as a right to credibility and reputation. Like, I don't know how else to describe it, but there's this thing that you have where you really under-index on what someone has done. In fact, I find in many cases, it's a knock because your concern is like, well, if you've been so successful, why are you going to come to glean and grind here? And working hard is such a core value for you. I don't know. Maybe tell me if you think I'm wrong. Yeah. I mean, I don't think I will take somebody's accomplishments to sort of start to think negatively about them. But it is true. I think hunger is a big driver for people to produce best work. And you have to be careful, like for somebody who's super accomplished, they may, like I do feel like, you know, that they may be less tolerant of like what they will have to go through at the company. Because they don't need to. They don't need to. And they'll question more. Like, do I really need to be doing all this? Do I really need to grind so much? So that is something that is, of course, you know, something I think about. The other thing is today, a lot of that experience, if that sort of becomes the right way for you, it has to be done that way because you've done it once and not twice and it has always worked for you in the past. That may not actually be the right thing for a company like us because I think the world is changing. Organizations are changing. How any given function is built has to change. The ratios of like, you know, do you need one solution engineer for one account executive or whatever those things are. Like none of that traditional, those metrics or those processes in some ways are all directly applicable anymore. Like you have to have the mindset to do things differently because a lot of work that you would need humans for before can be done by AI. So you have to have that open mindset. But there's no substitute, just to be fully transparent. Like ultimately, somebody who's achieved success, somebody who's seen a lot, they have something of value that others don't. Those learnings are invaluable. So I actually appreciate that. You just have to make make sure that, you know, that doesn't become the ceiling for them. Like, you know, they have to still be able to reinvent themselves. How are you using at home personally? Yeah. How are you, like, do you have any favorite things that you're doing with AI today? Like, how do you use it? You know, I got asked this question a few times and, and it sort of made me realize that I'm actually doing very little things outside of work. So you're like running, you're like figuring out what model you want to run on the backend for Glean and then doing new queries on it. So, So it's sort of the... How much are you working right now? I think I'm, like, I work pretty much most of my, most of the times I'm awake. So, like, I take some break, you know, like, and the break is to, the break is when I'm not trying to use more AI. Like, you know, I'm actually, like, trying to watch TV and relax and then go to bed. But the, so my AI usage has been largely at work. Right. And I feel like this year, like in the last, I would say three to four months, is when my own habits have fundamentally changed. Before I was using AI, of course I was using a lot of Glean, I was using ChatGPT, and I was using it mostly for knowledge seeking. I have questions, I want answers for those questions. I'm trying to find somebody to go and talk to on a given topic. So those are the kind of questions I would ask. But lately I've really transitioned to, but I feel like Glean is a more powerful colleague of mine than any other colleague. And I don't mean to, I'm not trying to be underappreciative of our great team, but there's something about AI that makes it a very effective personal colleague or companion to you. And so a lot of complex work these days that I have, I always feel guilty of I always feel guilty of like I have like my the list of questions that I have that keep coming to me my mind is never like sitting idle like it's just like racing and like it bothers me but I have just a constant stream of questions coming to me and I feel guilty like you know sending those questions out to my team because it's going to take time they're going to get distracted and I've not figured out that like actually like AI is incredible for that. Like you know, deep strategic questions and I can ask Glean to actually go work on it. And given that it has all that context of our company and how we are businesses, it's just incredible like, you know, the kind of things it can do. So like this is my new working model. Let's say I have a project, I need to make some strategic decision or I want to actually understand how, you know, we're doing as a business very, very deep fundamental questions, I will first start and ask Clean to do deep research and give me a report. A report that is two pages long that contains enough information for me to build a point of view. And once I've done that, that is when I start to actually engage with the rest of my team. And that way I'm actually coming in way more informed. I'm way more surgical in terms of how I actually consume or take their time. And in fact, many times, I will actually share the work of AI, you know, with the team so that I'm also trying to change their habits, like as opposed to them starting without AI. I actually share the AI artifacts and then, and sort of drive this behavior, but it's actually really incredible, like how like my work habits as a CEO has changed. And by the way, you eliminate a lot of the like bias. Like if you ask somebody in finance or marketing about something in finance marketing, they're going to think about their own worldview and then how they project that back to their boss. Yeah. Whereas at least now you can have an opinion that is relatively objective and press against that opinion. That's right. Comprehensive and unbiased. That's, you know, this is something that, these are like net new capabilities that AI brings. Have the questions in your mind gotten louder? Like, are there more or less than when six years ago? Like, do you have more hesitations, concerns, paranoia, questions that your mind racing? Is it getting worse or better? For us as a business, I think it has sort of, like initially we had a lot of anxiety because we had no success. And it felt like a slog and there was a lot of rejection. And so I had a lot of questions, but then we started to see success and we were clearly the only game in the town, the best product and all of that. And now we are in this place again where there's a lot of noise in the market. And there's a lot of existential risk actually to us. And that risk comes from the fact that if you step away for a moment, somebody will pass you. You have to be totally and fully alert. So it does feel that there are more questions on my mind now compared to last year. and obviously like more stress, but at the same time, I can also see that the opportunity is 10x bigger than what it was last year for us. Yeah. Well, I appreciate you doing this. It's been maybe one of the most rewarding things of my Kleiner Perkins career is being able to be close to the action with you and Glean. Like it, I'm telling you, when I went to that Glean user conference, it was a Glean user conference and it was full of people. Customers. I could not believe it. I think I pulled you aside and I was like, what is going, I cannot believe, like I could not believe it. Yeah. It's amazing. It's been a great journey and it's been a great conversation. Thanks, Arvind. Good to see you. Are you hiring? Are there any roles you're hiring for? We are, I mean, across the board. We're hiring engineers, a lot of salespeople, leaders also actually quite a few across. Like, for example, we're looking for a leader to run our federal program. So yeah, all the postings are online on our website, but definitely the companies are going to grow. We're likely going to double again this year. Wait till you see the 2000 number ahead of you. That's going to be daunting too. Thanks, man. I appreciate you. Thank you. That's it for now. If you liked the episode, please leave us a review or go back into the Archives, where we've done more than 200 episodes with some fantastic folks. This podcast is a Kleiner Perkins production, and I'm Juven. Thanks for listening.