AI Enterprise - Databricks & Glean | BG2 Guest Interview

In this BG2 guest interview, Altimeter partner Apoorv Agrawal sits down with Ali Ghodsi (Databricks) and Arvind Jain (Glean) for a candid, operator-level discussion on what’s actually working in enterprise AI—and what isn’t. They unpack why 95% of AI projects fail, why LLMs are rapidly commoditizing, and why durable advantage is shifting to proprietary data, agentic systems, and workflow integration. The conversation dives deep into real-world use cases across finance, healthcare, and retail; the debate over whether we already have AGI; and how AI spend, CapEx, and valuation bubbles will realistically play out. A must-watch for builder, and investors navigating the AI transition inside real organizations. Timestamps: (00:00) Intro (01:00) Consumer AI vs. Enterprise Reality (02:15) Why 95% of AI Projects Fail (04:15) RBC, Merck, and 7-Eleven Use Cases (06:45) What Actually Makes AI Work (07:00) LLMs Are Commodities—Data Is the Moat (08:45) Failed AI Bets at Databricks & Glean (11:00) RPA vs. Generative AI (14:15) Advice for CIOs Planning AI Budgets (16:00) AI CapEx and the Revenue Math (18:00) The Three Camps of AI (21:00) Making AI Useful Inside Enterprises (24:30) Why Apps Capture the Value (30:00) The Future of UI, Voice, and Data Entry (37:30) Rapid Fire: Winners, Bubbles, Long/Short Produced by Dan Shevchuk Music by Yung Spielberg Available on Apple, Spotify, www.bg2pod.com http://www.bg2pod.com Follow: Apoorv Agrawal @apoorv03 https://x.com/apoorv03 BG2 Pod @bg2pod https://x.com/BG2Pod

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  I think we have AGI. I think we have artificial general intelligence. We really have.

• 0:04Speaker 01

  You you hear these ninety five percent of projects fail, but like, you know, like, that's that's that's actually what you want.

• 0:09Speaker 00

  I I think the L L M is a commodity. People are not saying that, but it is a commodity. Like, you can get gas from this gas station, you can get gas from that gas station, it doesn't matter, just compare price.

• 0:17Apoorv Agrawal?

  Is AI in a bubble?

• 0:19Speaker 00

  There is an AI bubble. Okay, so then Lean is also in the bubble, everybody's in the bubble. No, I would I would say there is a bubble. I I would say those three camps.

• 0:27Apoorv Agrawal?

  Yeah.

• 0:28Speaker 00

  There is a super intelligence quest camp. I would be very worried there. There's a second, the researchers doing the, you know, that's definitely not in a bubble. They're like the.

• 0:36Apoorv Agrawal?

  They're sober.

• 0:37Speaker 00

  Yeah, they're they're super sober and nobody cares about them. And then there's, right? And they're probably the ones that are right, unfortunately. And then there's the third camp, which is us trying to make this valuable. We're not in a bubble in a sense that we're not spending huge amounts of capital on what we are doing. Uh, we're just trying to get actual economic value inside of these organizations.

• 0:59Apoorv Agrawal?

  Two legendary builders, Ali Irvin. I'm so thrilled to get into this with you because both of you have seen every super cycle I've lived through. Internet, mobile, cloud, data and AI. Not just through the super cycles, but also through the hype, the trough of disillusionment, and this time it's different. Today, we're gonna chop it up on the state of AI. You know, let's let's start with a twenty thousand feet view. Take stock of where we are. AI, we've seen consumer AI, billions of users, ChatGPT said the guns went off three years ago. Cloud perplexity, ChatGPT, people use it in the room.

• 1:41Apoorv Agrawal?

  On the SMB and developer side, you've got hundreds of millions of users with cursor and codex and cloud code and and so on. Enterprise, on the other hand, there's a lot of divide. It's hard to see a lot of fog of war. On one side, you've got models that are earning math benchmarks and science benchmarks and engineering benchmarks. But on the other side, you've got the MIT report that's saying ninety five percent of AI deployments don't work. What's the reality? Bridge the gap for us, lay it out as you see it, view from the top.

• 2:13Speaker 01

  So I think first of all, I think we should know that people use AI in their personal and work lives both. So there's not so much of a divide, like everybody in your company is probably using ChatGPT and Claude and other tools on a daily basis. The thing that I feel is happening in enterprises, you hear these 95% of projects fail, but like, you know, like that's actually what you want. Like when you are actually experimenting with new technology, if all of your projects are failing, that means you didn't just not trying enough, you know, at the moment. So I think when I read the study, like it was not a surprise for me. You know, we're going to actually see, hopefully, like, you know, similar stats next year too,

• 2:57Speaker 01

  because we want everybody in the industry to be really eager and experiment and actually figure out like, how to make, how to actually make, you know, get benefits from this technology.

• 3:08Apoorv Agrawal?

  This would make you guys by default be 5% of AI that is working,

• 3:13Apoorv Agrawal?

  which is 1 in 20. Maybe you go to the 5%. What is the use case that is working? And not just working like it's like saving me time, but like it's working and it's transforming my company. Something that you can take to the bank, to the CFO, while the CFO will notice it, but the legal won't shut it down.

• 3:32Speaker 00

  I mean, look, we're seeing a lot of use cases that are working. It's just that, you know, you just have to, it's not just you can just unleash the agents and it just works. It's an engineering art. Like if you're going to have a company that's going to be really differentiated, like my company or your company or anyone's company, and you want to beat the competition, you can't just like, you know, quickly put something together and think that, you know, that your competition is not going to do the same thing. So that's going to be, you know, something that needs evaluations. It needs something that, you know, you're going to productionize. It's going to take effort. You need a great team around it, but we're seeing a lot of them. Like I'll give you some examples. Royal Bank of Canada built agents with us that basically take, as soon as an earnings report comes out. So equity research analysts, their job is

• 4:16Speaker 00

  Is to put together these, you know, reports that say like, you know, this is a buy, this is, you know, hold and so on. The agent goes, gets the earnings report, gets all the previous earnings reports, get all the computers earnings reports, gets everything that's going on in the market, does the full analysis, the news, everything, puts it all together and it can get the equity report out in fifteen minutes from the earnings call. Industry standard is two hours. Of course, it's gonna get commoditized and others are gonna do that as well. But that's actually a really important use case that we're seeing in finance. So that's like finance, example in finance, right? It's like, and there's lots of examples like this, sifting through hundreds of thousands of documents, SEC reports, so on. That's finance. Let's switch gears, let's go to healthcare. Healthcare is completely different.

• 5:01Speaker 00

  In healthcare, we have a, you know, customer Merck that in the life science space created a model called TEDDY. TEDDY stands for Transformer Enabled Drug Discovery. And this is a transformer model, kind of just like large language models that can predict the next word, but it instead can figure out which genome is missing if you remove a genome. So it really understands the gene regulatory network and can really start telling you what's happening with gene expression and so on. So this is really important for drug discovery. It's the beginnings, but this is gonna actually help us do things that we couldn't do before. Let's pick retail also, so I'm picking different. Healthcare is one, I gave you finance, right, the RBC one. Let's go to retail, seven eleven.

• 5:45Speaker 00

  Agents that completely automate the marketing stack. Actually, the marketing stack is gonna get disrupted pretty heavily. So these agents can basically prepare, they can segment the audience, like this segment wants to hear this, and it can prepare all the marketing material that's like directly targeting you guys, and it can put the campaigns together and do that. Seven eleven was doing this before as well, but, you know, this, and we're seeing this at Databricks as well, more and more is being done by agents and being automated, so you can just do it faster and you can segment more fine grain. Because before you had to create the content for the groups, that was a heavy, content creation was something that was human manual labor. Now you can actually do that much, much more. You can have all your web materials completely customized for a target group. So these are examples where it is working.

• 6:30Speaker 00

  There are also lots of examples where it's not working, even with Databricks, we're not just the five percent, you know, we have that, some of that ninety-five percent too, but some examples of where it's, where it's, where we're seeing success.

• 6:40Apoorv Agrawal?

  Ali, follow up on that. These are great examples, thank you. Maybe if you were to take it a layer up, what is common across these use cases, or these organizations, or these CIOs that's making these use cases work? Is there something that we can pattern match?

• 6:54Speaker 00

  Yeah, look, I think the LLM is a commodity. People are not saying that, but it is a commodity. Like, and, you know, when I took econ classes, commodity was when it's interchangeable. Like, you can get gas from this gas station, or you can get gas from that gas station. It doesn't matter, just compare price. LLMs have become that way. Like, it doesn't really matter. This one is better right now. Next week, that one is better. You can't even keep up anymore, right? What's happening? So they're a commodity. So it's not about that. It really comes down to your company. What data does your company have that's special that your competitors don't have? Can you leverage that? And can you build AI that really understands that data? Because that's not a commodity. There's not an AI out there that understands all your business processes in your company, your secret sauce, and your data. That's not a commodity.

• 7:37Speaker 00

  In fact, that's closer to the 95%. It really comes down to that. Or if you have a complicated process that just your company has, this is how you deliver your product and services in your company. And it's, you know, it's... That portion can be disrupted with AI somehow. If you can do that, now you can get ahead of your competition. But it comes back to what makes your company special. Unfortunately, a lot of companies are just building commodity stuff. Like, you should not be building that because it's a thing that every company can do. It's not special to your company or to your... That's, I think, the problem in a lot of the industry. Another problem in the industry is a lot of demo ware. It's really easy to make cool demos with Gen AI. And, you know, therefore, we're seeing a lot of cool demos. But that's all they are.

• 8:19Apoorv Agrawal?

  Yeah. Well, you know, something we say around at Ultimate quite a bit is your AI strategy starts with your data strategy. And so you've got to get the data house in order first. And, you know, there's a lot of reasons for use cases that are, you know, we're trying, we're not working. Maybe give us an example of the 95% of an AI bet that either of you had, add Databricks, add Glean, that did not work out, and why it didn't work out?

• 8:43Speaker 01

  It's actually an interesting thing, you know, with engineering today is you build systems and and never never before have you been in this mode where you start with a great idea and it doesn't seem like a good idea anymore, like within two weeks. You know, because we see a new development that happened. So there are, we have like numerous failures failures in in engineering on that front. Like, you know, for example, some of our fine tuning work, building, building models for a specific use case within our product, like, you know, didn't really pan out for us. And ultimately, the choice was that, you know, we can go with already built models, whether they are small, open source models, host non data bricks, or or one of the large, you know, foundation models. The but internally, like, you know, from a corporate, you know, use cases perspective,

• 9:28Speaker 01

  actually, like, you know, we are also like in in many ways in this mode where a lot of our work actually, like, I would not say like fail, but it actually takes much longer than, you know, to actually generate success. You know, there are, you know, we are actually trying to automate a lot of our business processes internally. And like, for example, like, you know, one thing that I want is in our company, I want everybody to actually know exactly what their top priority for the week is, what they want to work on. And and maybe, you know, we want an AI agent to actually first tell them what their priority should be. And we want it all to be documented. And we want a system which actually then, you know, rolls it rolls it all up. And I get a view every week where I can actually

• 10:12Speaker 01

  quickly see, you know, what are all the different people working on in the company? And are they aligned? Is that aligned with, you know, what I want them to work on? And and this is a simple thing, like, you know, companies have always tried to actually have this, you know, as CEOs, you always want it. And it's always hard to make happen. And we thought that AI would simply just like, you know, magically do all of this work because, you know, like, it has all the context, really has all the context inside the company to make it happen. But I still don't have it. So so so things do take time to actually, you know, to Ali's point, like, you know, there is AI is just one more tool that you have in the toolkit. It does not suddenly make building complex enterprise systems, you know, you know, like, it doesn't make it like that you can, you know, build it up like in in one day, doesn't it?

• 10:56Apoorv Agrawal?

  Yeah. You know, the last time enterprises got this excited about a tool was called RPA. And we know how that ended. It unfortunately fizzled out. And, you know, somebody in the audience yesterday is like, hey, how is this time?

• 11:10Apoorv Agrawal?

  Different from RPA. It seems like the same movie, bigger budgets, better actors. What's different this time? How is the nature or the architecture of the technology different from the previous automation cycle? Either of you?

• 11:23Speaker 01

  Yeah, well, I mean, first of all, RPA, like, it didn't take, you know, it didn't capture my attention at all. So I've no, so I actually can't, you know.

• 11:30Apoorv Agrawal?

  What is RPA?

• 11:31Speaker 01

  So, so, so I think, I, I think, like, I, I would not compare these two technologies at all. Like, you know, you know, what we, what we're seeing now with AI is so fundamental. You know, it's, it's, you know, it's, it's the, uh, you know, when, when we saw it first, it was basically magic. And, and we couldn't believe that this is a machine that is doing this work. Machines just simply cannot do these kind of things, you know, that we saw them do. Like riding on their own, um, having emotion, understanding emotion. Um, so it's a, um, it's, you know, it's, it's fundamental, it's different. And, and the, uh, and, and that's why, like, you know, I don't think, you know, we, this, this, uh, this technology is going to fizzle out.

• 12:15Speaker 01

  Um, and it's not like, you know, you don't have to be like a financial expert or, you know, a, you know, like sort of a deep thinker on business. This is, this is obvious stuff, like, you know, all of us know, all of us feel it. All of, all of us can see the capability of this technology. And we know it's special and it's going to be, and it's going to be around.

• 12:32Apoorv Agrawal?

  Yeah.

• 12:33Speaker 00

  Do you want to hear my RPA?

• 12:34Apoorv Agrawal?

  Please.

• 12:34Speaker 00

  You know, I mean, it was rule-based. And the problem with it, especially if you're, you know, you're, you want something that automates what's going on on your desktop and automate the work that's happening. It's just that there's too much unexpected things that happen and it's just hard and brittle to set it up. It wasn't learning ever. So there was like zero learning. It was like, you tell it exactly here are the rules. And if it got something wrong, you, you need to go and go back and expand the rules. Here, you have something that's learning, right? So it can, it can improve and it can generalize and it can understand the patterns and do pattern recognition. Uh, so that's the fundamental difference between these two. Now, uh, there has been many startups that have failed in the generative AI. We're going to replace RPA with generative AI models.

• 13:16Speaker 00

  There's many startups that failed, actually, that I know of, like pretty some high-profile ones. It's because, um, the paradigm we live in today with AI is there's still problems. The biggest problem is that you bake a model. And that's where it's learned, everything it needs to learn, and then you freeze it. And then you launch it, and then maybe you give it some context, but that's it, it's frozen. So therein lies the problem, uh, that, you know, we need, we need an AI that really can sort of continue learning while it's using the desktop and clicking around. So I do think this problem is hard to, to, to nail, but I think Arvin is right, that it's like, there's no comparison at all. It's like rule, brittle, rule-based stuff versus a learning agentic system. Uh, I think it's gonna nail it perfectly, but we haven't really nailed computer use yet.

• 13:59Apoorv Agrawal?

  Yeah, working on it. The number one shift is this move from if, then, else statements to a more generative solution that figures out the solution. Um, and so you're trading breath for, for maybe determinism. Um, that seems to be the difference. And, you know, there's a lot of CIOs in the room we've got here, and they've got budgets coming up to plan. Uh, if you were giving advice to them of like, hey, based on everything I know from my customer base, here's one thing or two things that you've gotta figure out and, and align incentives on, or it could be reliability problem or org design. What advice would you have for CIOs who are thinking about their AI budgets right now?

• 14:37Speaker 01

  Well, spend more. Spend more, yeah, put it on lean, but the, I, I think the, um, like, one thing, you know, which, uh, which is important in AI market today is that it's very new and there are many players. In fact, um, every, every software company is also an AI company now. Um, you can go and check their websites. So, so the, I think it's, it's just hard to actually figure out where to allocate those budgets. And, um, what we tell people is that I think the winners are yet to be identified. And, and so experiment with more vendors, do short-term, shorter-term contracts.

• 15:19Speaker 01

  Um, and, and, you know, while that's easy to say, uh, it's hard to actually implement, because every product that you try has, you know, it's a cost that you have to pay to make it, make it sort of even tested. So, so you had to also pick products that are, are easy to test. I mean, those are the, the ones that don't require you to, you know, spend the next six months trying to implement something, and you have no idea what's gonna come out after that. Like, you know, the products of today, the, the products that are built with the right AI, they should work, you know, um, very, very quickly for you.

• 15:51Apoorv Agrawal?

  Crawl, walk, run. We're going to take a peek into the future, shifting gears. You know, one of the things that keeps investors like me up in the nights is a quarter trillion being spent on Nvidia on the semi side of things. Assuming that is just 50% of the CapEx, you're spending about half a trillion on CapEx. And then you've got to earn about a trillion dollars of AI revenue for all of this CapEx to be worth it.

• 16:16Apoorv Agrawal?

  This, and just to put this in context, the entirety of the software industry earns about $400 billion of revenue.

• 16:24Apoorv Agrawal?

  This seems like a physics problem at this point. How do you, how do you think this, this plays out? You know, you've got, you've got to make about a trillion dollars of revenue to justify this present spend that's already happening. How do you think this shakes out? Maybe we start with you, Arvind.

• 16:42Speaker 01

  Wrong person to start with, but you know, I'm an engineer and I actually don't really, you know, think too much about who's spending what money. Like, you know, we're here to build our product and add value. So that's, so in some sense, you know, I've not really thought too much about this problem. But, but if you think about AI, the, you know, AI is not actually, you know, extending software in a marginal way. It's a, it's a, it's a different product. And in fact, you know, it's actually going to grab a lot of revenue that actually today is in services industry, which is 25 times larger than software industry. So there's, there's a lot of spend that is going to move. I mean, the spend that you see happen on AI is actually sort of, you know, those service dollars that are converting into AI or software dollars.

• 17:24Speaker 01

  And, and I think the, but with that said, you know, maybe, maybe you have a more informed view on this.

• 17:31Apoorv Agrawal?

  By the way, I do think that's, that's just to build on, you said, you know, I'm in engineering. I want to just build something that's cool. I, I do think it's not binary, right? It's not like, okay, so the physics doesn't work out, so the whole thing will collapse.

• 17:41Speaker 00

  No, there's gonna be things that work. And so it is a good idea to continue focusing on the stuff that is obviously already working. Continue expanding on that. But I think if you zoom out, I think there's like three paradigms or three kind of camps. And I put Arvin in the third camp. I actually put myself also in the third camp. But let's start with the first camp. I think the first camp is this quest for superintelligence camp. And it's, you know, I think all the frontier lab labs are doing this. Like in all three, four or five of them, however you want to count them. And I think it's really still being, a lot of it comes from the scaling laws mentality. Which is whoever has the most GPUs and the most data is gonna win the quest for superintelligence,

• 18:26Speaker 00

  which is kind of intelligence that's like on almost like godlike. It leads to recursive self-improvement of the AI, which then once you have that, it can cure cancer and solve all economical problems. And we can probably 10x GDP over a few years period of time. So what the hell are you talking about? That there's a physics problem. Like anything, any of your cost equations are gonna pale in comparison to the economic value that this thing is gonna provide. So that's like one camp. And the way they're developing it is bigger and bigger clusters, more and more energy. And that's how they're going about it. And that's where most of the capital is going. Right. Right? That's not you, that's not the kind of capital you're spending or I'm spending. But that's that camp. And then how do they know that they're succeeding? They're not just like, oh, just trust us. They're, you know, very smart people working on this.

• 19:10Speaker 00

  So the way they're approaching it, they're saying, you know, we'll throw the hardest questions we have at the whatever AI we have now. And if it nails them, and we're making really rapid progress, so what's your problem? Like, we're like, look at math Olympiad. We're like nailing these math Olympiad problems and physics Olympiad. And programming contest is like better than any human being. So like, that's what they're doing, throwing all the most intellectually challenging. There's a second camp, which are the people that created the original technology, the scientists who created the technology, got the computer science Nobel Prize for it, it's called Turing Award. And that's, you know, Rich Sutton, who created reinforcement learning, which, you know, a lot of this stuff is built on. You have Jan LaKoon, who was one of the three founding fathers, and many others. They have for many years, actually, I've been, you know, I've asked them for years,

• 19:54Speaker 00

  they've been saying that that first camp is not gonna, that's like not even the right approach, is their view. They're like, no, that's just like autoregressive next token prediction, it's just probabilistically predicting the next token. That's not how, and usually they will say, that's not how humans learn, that's not how animals learn, you know, that we operate in a different way, your brain is not that way. And one example is that, you know, even a child learns very quickly to walk and talk and do things with very little data compared to, you know, like certainly no child is reading all of the internet's data four times over before they learn to speak. So I don't, so that's like camp number two. Those guys, by the way, they say it's 20 years out. So they're saying, hey, it's a physics problem and it's gonna take 20 years to get there,

• 20:37Speaker 00

  which to me it's like, I don't know, like, leave me alone, let me do research. Third camp, which is, I think, what we are in, is I don't think we need superintelligence, like, you know, I don't think we need that superintelligence right now. Maybe they'll get there, that's awesome if they do, but I think we have AGI. I think we have artificial general intelligence, we really have it, we absolutely have it. It's like, anyone who says we need to get to AGI, that's like, it's, it's a, it's false premise to start with, we already have AGI. I came to the United States in 2009 at UC Berkeley, not far away from here, and I was in a AI lab, it was called AMP Lab, the A was for algorithms and AI, machines and people, and these are all AI people, and back then, the definition of AGI we had, we already have satisfied that.

• 21:19Speaker 00

  Like, I know the discussions we had, and I actually went back to some of those folks to see, like, is it just me, or what was the sentiment back in 2009? And everybody that I talked to said, yeah, that's by those standards, we had AGI, but we've changed the definition now. We have those definitions, you know, adds. So for 30, 40 years, we had a definition of AGI, we've already hit that. Now we're changing it and moving the goalposts. But very obviously, we already have AGI. Just use any of these L L Ms and have it do some reasoning. And certainly, it's smarter than many, a lot of friends that you have that, right? Like, you know, let's, let's not them or coworkers or whatever, right? So you already have AGI. Now, now we're like haggling over exactly how smart is it? You know, do you have a friend that's smarter or not? So if we already have AGI, we just need to make it useful inside the enterprise.

• 22:04Speaker 00

  We need to just expand that five percent to be 10%, 20%, 30%. So that's why I think our answer is actually a good answer. Like, we have the AGI we need. Let us just focus on solving the actual problems inside the organizations. And I think we can already, that's, that's enough to automate a lot of the tasks and get huge economic value out of it. We don't actually need superintelligence for that. That's a good idea. If the superintelligence guys nail it, amazing. Then we've cured cancer. If they don't, hopefully, the second camp comes up with a new thing in the next 20 years. That's also awesome. We already have whatever we need. So, yeah, yeah, let us just do our engineering.

• 22:40Apoorv Agrawal?

  Right. Yeah, that's really good framing. And the way this manifests in the, in the, you know, in the world is there's a data layer. There's the intelligence layer, which is where Camp One is presumably producing a lot of great models. And then there's the software layer where the users engage with.

• 22:57Apoorv Agrawal?

  Where do you think value will accrue? If you were to design 100 units of value across these three layers? The data layer, the intelligence layer, and the software or the application layer? Where do you think value accrues in the next five years?

• 23:11Speaker 01

  All right, this is a, this is a tough question. I mean, I think the, all, all those three layers actually are very fundamental.

• 23:18Apoorv Agrawal?

  Yeah.

• 23:19Speaker 01

  I thought you were gonna add a few more, which are not.

• 23:21Apoorv Agrawal?

  Yeah, we could have gone down to info and semis, but.

• 23:23Speaker 01

  Yeah, because I think, I, I feel like the, like, as, as Ali was saying, that the models are going to be available to all of us. You know, they are going to be commodity. It's going to hard to sort of see that it, the more spend goes to them versus, you know, these layers on top. Um, the but how do you how do you like, you know, I it's hard to sort of come up with, you know, where the most value will be. And and I also don't know if if actually it changes from today's technology architecture, where again, like, you know, when you think about in a pre-AI world, uh, any any sort of like, you know, enterprise, um, you know, uh, you know, a application and data systems, you know, you have you have data systems, you you do have, I guess you you don't have enough of

• 24:08Speaker 01

  that intelligent layer today, and then you you have the application layer. So so I so I so I guess, you know, some some dollars will shift into it. I know we do think that in in the intelligence layer is actually going to be pretty thick one, maybe, you know, it's it will capture half of the enterprise value.

• 24:21Speaker 00

  Anything to add, Ali? Yeah, no, I I I think that, you know, uh, yeah, there are more layers in the stack, depending on how you want to do it. But um, I think, as I said, the the L L M is a commodity, uh, as Arvind said, you can get them like, you know, but that's not that doesn't mean those companies are not going to be valuable. They can be very, I mean, TSMC is very valuable, but um, I'm saying, and they're going to be kind of like these fab-like companies, but uh, but they're interchangeable, and we've never seen something like that ever. I I have not during all these, people just switch L L M's like in one day. That's not the case with your, you know, your iPhone versus Android, or your Windows versus your Mac, or your anything versus anything, like, you know, uh, you know, Google Sheets versus Excel is like huge religious battle inside our company.

• 25:04Speaker 00

  But but L L M's, it's like, you know, because um, it's a commodity, as I said, it just speaks English or any language you like, and you can, and it gives you different answers every time, might as well just try the cheaper one, the cheaper commodity, or the flight, the smarter commodity, you can't even really tell the difference, can you? So then what is special is the data that you have. Again, if your company has data that it has actually collected, that your competitors do not have, like, Glean is amazing, but if you remove all the data from Glean, it's there's no use to it, right? So it's all about the data that you have. And can you secure the data also? So if we're going to have agents running around accessing this data, like, oh, that's his H R data, oh, here is the poor was salary information, oops, I blurted out to all of

• 25:47Speaker 00

  you, like, you know, like, that's, you know, so how do you lock it down? Uh, how do you make? Sure that um the there's governance, there's you know, there's also a lot of worry around can What if it's using a Chinese model? What if it's accessing this information? What if it's sharing this information with a competitor? What if it's interacting? So the governance, security layer is going to be super, super important. But I do think most of the value will accrue to the apps.

• 26:09Speaker 01

  Yeah.

• 26:10Speaker 00

  So it's kind of, and I think that's common sense. It's just, I just don't know which apps.

• 26:14Speaker 01

  Yeah.

• 26:14Speaker 00

  Uh, I do think Lean is amazing. I don't think, do you think of it as an app or I don't know, but.

• 26:20Speaker 01

  Now we see us as both app and a platform, but yes.

• 26:23Speaker 00

  So I think it's a, that's called an app platform. I do think it's amazing.

• 26:26Speaker 01

  Yeah.

• 26:26Speaker 00

  Because it has the potential to automate uh so much of the overhead inside of an organization. Like if you think about, why do organizations have hundreds of thousands of employees? You know, some organizations or fifty thousand, twenty thousand. A lot of it is the coordination overhead of like, you know, so many people have to communicate with each other. Hey, what happened? What did you exactly mean by this? Let's do a meeting where you explain to me. I ask some questions. Let's, or let's invite these other guys also and then write it down. And then just just the coordination overhead of organizations is massive, right? It's like this N squared problem that, you know, everybody needs to communicate with everybody and they're communicating inside their siloed org chart, but how do we get it across? So this, like, you know, through docs and Excel sheets and PowerPoints and meetings is how we like move companies and organizations forward.

• 27:08Speaker 00

  So much of that can be augmented and be made more efficient with Lean. So that's why I think Lean is amazing. But this is kind of like uh two thousand.

• 27:18Speaker 01

  Yeah.

• 27:18Speaker 00

  And you ask, what are the killer apps on the internet? By the way, back then, we thought it's like Cisco routers, you know, portals maybe with thousands of links on them. Actually, I was like in, I just started college and that we knew that the future of internet would be portals, which are these webpages with a hundred links on it. And you just quick click on the right link. This is before Google search. But the fu- future of internet actually didn't look that way. It ended up being, you know, things like Facebook for friends and things like Airbnb for rentals and Uber for your cab industry and, you know. Twitter and so on, those became great companies. So I don't know what those are for the future.

• 27:53Apoorv Agrawal?

  Yeah.

• 27:54Speaker 00

  They will pop up.

• 27:55Apoorv Agrawal?

  Yeah.

• 27:55Speaker 00

  And they will be extremely valuable. But, okay, so does that mean that Databricks and Glean then basically will die? And there'll be a new set of companies. No, back then, there was actually an Amazon.com already in ninety eight. There was already a Google actually existed already in ninety eight and so on. Cisco, by the way, is still around and it's, you know, only a three hundred billion dollar company or something like that, right? So, uh, so it's not binary. We'll see what happens. But I do think there's gonna be really, a lot of value will go to the future.

• 28:21Apoorv Agrawal?

  Yeah.

• 28:21Speaker 00

  You know, apps that will emerge.

• 28:23Apoorv Agrawal?

  Speak, let's double click into that.

• 28:24Speaker 00

  Yeah.

• 28:25Apoorv Agrawal?

  The three hundred billion dollar companies of today add that layer, software, apps, Salesforce, ServiceNow. Lot of talk about software as being dead. Satya calls them the crud apps. What is the future of this layer that today is called software that seems to be heading towards becoming a database? And what do you see the, the, the, the value accrue to, to, to those, to these, uh, this part of the layer? Maybe start with you, Arvind.

• 28:52Speaker 01

  Yeah, I, I, I think that's a oversimplification. Like, for example, even to say that Salesforce is, um, you know, is just a database. Um, you know, it's a, it's a full sort of ecosystem of workflows. And other applications, you know, that are sort of built on top of that infrastructure. So.

• 29:14Speaker 01

  So I sort of like, you know, I haven't really understood this concept of that, you know, you have this. Like a, you know, database, you know, where all your enterprise data is and then. And then people can just go and create dynamic UI experiences on their own, on top of that data. On it like, you know, every business can, for example, just create all the UI by themselves on this. I don't, I don't think, you know, it's going to be. Happening like that because, yes, you know, AI makes it easy for you to build. You, you can have a database and you can build, uh, you can just talk to AI and create a UI and experience that. That, um, that is, you know, exactly what you want it to be.

• 29:54Speaker 01

  But most times you actually won't know what you want. Like, you know, I think a lot of like, you know, good thing about software companies is that they actually think about. Uh how to actually take that data, but then uh present it in a way uh let you know make people interact with it or modify it in a way. Which sort of is natural and which you know drives you know like more productivity from a human. So so I think ultimately like a software is an end to end stack in my opinion and. Uh all of these companies, you know, I don't think they're going away. I don't think you know they're gonna relegate it to becoming a database.

• 30:23Apoorv Agrawal?

  Humans, you know, over the last twenty years, we got addicted to these screens. We scrunch over the screens and we would input this information with their with their keys. With the dropdown and here I met Irvin today and this is what I learned. It should really be, hey chat, met with Irvin, this is what I learned. Remind me in two days to catch up with him. Right?

• 30:38Speaker 00

  That will happen. That I think is gonna happen. Yeah, that will happen in the next couple of years and even Glean, you you won't be wanted to type, you wanna talk to it. But I think the big thing is data entry.

• 30:46Apoorv Agrawal?

  Hmm.

• 30:47Speaker 00

  How does the data appear in that database?

• 30:49Apoorv Agrawal?

  Hmm.

• 30:49Speaker 00

  And that's today uh not completely automated. So you know, just just like uh I I think a company that would be well positioned to do that would actually kind of be Zoom. You know, and a lot of people don't think about it that way. But Zoom is really should be the the the perfect data entry uh application, right? Because that's where you're having all the conversations and that's all the information's coming out. And if they could, you know, if it could work with Glean and get extract the most important information.

• 31:15Apoorv Agrawal?

  Yeah.

• 31:16Speaker 00

  Store it all, not in like a structured day table, but like store that information in system of record.

• 31:21Apoorv Agrawal?

  Hmm.

• 31:22Speaker 00

  If you had that, that would be the full disruption of the SaaS.

• 31:25Speaker 01

  We have that already. That's that's actually is is is one of the most common agents these days, you know, with Glean, which is you take these meeting recordings. You figure out like, you know, what you talked to the customer, uh what were the action items? And then the agent goes updates the notes, you know, sales force with that, like these these kind of things are happening already. Yeah. Meeting meetings is, yeah, is uh. I you know, like we in in Glean, we have this uh um policy where we record every single meeting, internal meeting, external meeting if our customers allow. Uh, cause there's so much, so much information, you know, in there.

• 32:00Apoorv Agrawal?

  I, I joined a meeting last week. It was four humans and six AI note takers.

• 32:05Speaker 01

  Yeah, I heard about, I think yesterday we were talking about seventeen note takers in one of the, you know, discussions.

• 32:10Apoorv Agrawal?

  This, it felt like the first, you know, the first scene of a movie where, where the AI takes over. Clearly, there's a lot of sprawl. There's, there's like only, almost too many tools and consolidation coming.

• 32:20Speaker 01

  Yeah.

• 32:21Apoorv Agrawal?

  At some point, but, but maybe your guys' personal workflow, you know, you guys are CEOs in the age of AI. A lot of CIOs in the room, they've got more jobs than time on their hands. How are you using AI for both your personal self? And how are you driving your organizations, both large organizations to, to adopt AI and, and benefit from it?

• 32:40Apoorv Agrawal?

  Maybe give us a glimpse of your leadership in, in the age of AI. Maybe I'll, we'll start with you this time.

• 32:44Speaker 00

  Yeah, I mean, we have agents for all kinds of stuff that we use, you know, everything from, you know, we have agents that are really good at understanding our customers. We have an agent, Refi, Refi is the name, which, you know, yeah. If I want to understand anything about any, like, you know, tell me best customer story on this. Like, you know, I told you about RBC, Royal Bank Canada, but I can just ask it. I need a use case. I'm going to get on stage. I need to talk about finance sector. Give me a use case that has these, it'll like just find you all the information collected. So it's really, really helpful for me for these kinds of things, like when I get on stage like this. But also customer, if you go into a customer meeting, you know, I want to tell customer X about their biggest competitor Y, how they're using Databricks.

• 33:25Speaker 00

  Now, maybe Y is not a competitor, is not using Databricks, so then I shouldn't use, I should use Z, which actually is using Databricks. Maybe that's like the number two competitor. How do I get this information super quickly? All of those are prepared at Databricks. So on the go to market side, a lot of this is being completely automated and we're using this. The marketing stack I already mentioned is heavily automated already, like the, a lot of the tasks that's happening in marketing. So we're seeing that stack that's happening. Then there's engineering. That's like a whole big thing. Like, you know, that's how we sort of, and I think there's a whole change management and how to do it right. You know, initial attempts at automate a lot of the software engineering at Databricks kind of failed. Even, there's nothing wrong with AI. The problem is the humans and how we were organized.

• 34:10Speaker 00

  But that's, you know, so those are like the two big orgs. Databricks is a big, you know, six thousand person go to market org and three, four thousand person R and D org, and then there's some back office stuff. Those two already, we're seeing heavy automation using agents for all kinds of the tasks. Then there's a back office, so there's a finance and these functions. Finance is all on Databricks and it's all, all the forecasting, all the sort of, it's all moved to machine learning based. But it took them a long time because they had their Excel models and they're very proud of them and they didn't wanna, you know. But uh, um, again, there's a change management there. We actually had an external data science team build the AI models. And then eventually they became good enough and now uh, finance has taken those over and like now uh, finance have kind of moved from Excel to Python.

• 34:56Speaker 00

  Uh, largely at Databricks, but it was a journey because you know, most of us speak, we speak Excel. Similar thing is now happening to HR uh, and other departments as well, but I think they're like, you know. I think in general, HR departments are like, you know, even like they're not the closest to doing this kind of analytical work with, you know, Excel and so on. Uh, so may- maybe that's not quite as far along, but yes, it's, we're seeing it everywhere. Yeah. Anything to add, Arvin?

• 35:22Speaker 01

  Same for us and I think I can share some of my own personal use uh, with, with it like so. So one of our agents is daily prep agent, which I really love because you know, every morning um, it tells me like, you know, what my day is going to be. What I need to read, what I need to prepare, like most of the meetings, you know, I will not have context, it actually brings, you know, like the plan for those meetings for me. So that's, that's one of my favorite um, agents, you know, that helps me feel more confident like, you know, like how I'm gonna do my meetings in the day. Um, the other one uh, which I, which I shared yesterday also um, the like, you know, I, I've changed my instinct and I think you know, changing, changing instincts, you know, take, take, take, take a long time.

• 36:03Speaker 01

  Um, and you know, when, when you're the CEO, like you're the boss and everybody listens to you and you can just like, you know, whenever you have, you know, a small question, curiosity, just go and ask somebody. And they're gonna like, you know, uh, put thirty people on the task to actually get that, get that answer for me and this is.

• 36:17Speaker 00

  They're gonna have a prep meeting before the prep meeting.

• 36:20Speaker 01

  Exactly, yeah, so, so it's all of that, so, so that, so that's sort of like, you know, and, and so, but it's sort of like, for me, it was easy. I just get to ask somebody and, and that, you know, uh, I changed that because I knew I was actually causing like, you know, a lot of um, that was, that was very expensive. So the, so today, like, you know, my, my instinct is to, whenever I have curiosity, whenever I have questions, when I need to do. Uh data analysis when I need to write something, you know, my my little letter to the company every month. All of those things, you know, like fundamentally, I use, you know, AI, of course, you know, clean in this case, but to actually help me. Um, um, do my, do my tasks.

• 36:57Apoorv Agrawal?

  Yeah.

• 36:57Speaker 01

  More, more, I think the uh you have to, you you have to sort of have that. Um, belief, a lot of people won't do it. You have, you have to have that belief that AI is a good collaborator. It's not gonna do the work for you, but if you use it, you can actually produce better output eventually. Even if you don't save time, you know, for the first, you know, first few months, but you're actually gonna improve the quality of the output.

• 37:21Apoorv Agrawal?

  Fascinating. Well, this brings me to my favorite part of this conversation, which is rapid fire. Short answers are fine, long answers are welcome. Um.

• 37:33Apoorv Agrawal?

  Start with the twelve months from now. Are the big AI companies that we know of today up or down? We'll start with OpenAI, twelve months from now, stocks up or down? Ali and then Arvind. Up.

• 37:47Speaker 01

  And I'll say revenue will be up. I don't really understand how stocks work.

• 37:52Apoorv Agrawal?

  Anthropic. Ali, Arvind. Up.

• 37:55Speaker 01

  Same, yeah.

• 37:56Apoorv Agrawal?

  Okay. Are we?

• 37:58Speaker 00

  Let me, let me, can I give you a little bit more?

• 37:59Apoorv Agrawal?

  Of course.

• 38:00Speaker 00

  Because ChatGPT is gonna continue growing and it's on fire and it's what everybody uses.

• 38:04Apoorv Agrawal?

  Hmm.

• 38:04Speaker 00

  Uh, so is Gemini, by the way, and then Anthropic, because more and more, you know, coding, we've only like eaten into a small portion of that market, it's just started, so.

• 38:13Apoorv Agrawal?

  Yeah. Is AI in a bubble? Yes or no?

• 38:18Speaker 00

  There is an AI bubble, uh, like saying, like, okay, so then clean is also in the bubble, everybody's in the bubble. No, I would, I would say there is a bubble. I, I would say those three camps.

• 38:28Apoorv Agrawal?

  Yeah.

• 38:28Speaker 00

  There is a super intelligence quest camp.

• 38:30Apoorv Agrawal?

  Mm hmm.

• 38:31Speaker 00

  I would be very worried there. There's a second, the researcher is doing the, you know, that's definitely not in a bubble, they're like the.

• 38:37Apoorv Agrawal?

  They're sober.

• 38:38Speaker 00

  Yeah, they're, they're super sober and nobody cares about them. And then there's, right? And they're probably the ones that are right, unfortunately. And then there's the third camp, which is us trying to make this valuable. We're not in a bubble in a sense that we're not spending huge amounts of capital on what we are doing. Uh, we're just trying to get actual economic value inside of this organization. So I, I don't think it's binary, but there is a bubble. I mean, there are startups with zero revenue worth, you know, ten, twenty, thirty billion. That's a bubble.

• 39:05Apoorv Agrawal?

  Yeah.

• 39:06Speaker 01

  Same. I mean, I think the, uh, there are quite a few companies where there's all optimism and valuations, which are well ahead of the business that those companies have. And like, I guess you can say like, you know, compared to non-AI companies, like of course, AI companies do have, um, higher, higher, higher multiples and, um, but I think, you know, this sort of comes from that, you know, that, um, there's a good reason for it, you know, because, you know, these are, these AI companies are going to grow, uh, more than non-AI companies, for sure.

• 39:36Apoorv Agrawal?

  Yeah. My favorite game at Ultimatum, we ask our CEOs is a long-short game, is if you were to pick a company, a product, an idea, that you're long, that you think is going to be a bigger deal than it is today, what is that? And then short, which is, you know, there's more sizzle than there's steak, more, more hype than reality. Pick a long, something that you're very optimistic on. Same order, Ali and then Norvind.

• 39:57Speaker 00

  I am very long on agents. You know, I think I'm very long on speech.

• 40:03Apoorv Agrawal?

  Speech.

• 40:03Speaker 00

  As an interaction, like I think keyboards are kind of basically going to disappear completely. We haven't actually nailed speech. I know, I know it feels like we have, but we haven't because you're still using your keyboard. So as long as you're using a keyboard, we haven't nailed speech. But I think we're this close to completely eliminating, uh, uh, keyboards. So I think that's, that's a big one. What's, what would I say? It's like, you know, I do think coding is a little bit overhyped. I don't know if I would short it. It's, I mean, I think it's still the future. So I think that's, that's one of them. I think automating our customer service and support is a little bit overhyped. So, you know, I basically, I think the things that the industry, things are like amazing and we've made great progress. We probably haven't done as much progress. And then, and then a lot of the other things that are being ignored, uh, you know, we're going to have breakthroughs in those, so.

• 40:46Apoorv Agrawal?

  Fascinating.

• 40:47Speaker 00

  Yep.

• 40:49Speaker 01

  Yeah, and for, for me, I think the products that are, uh, going to change the paradigm, uh, where instead of you building a product and, you know, expecting people to come to you, um, if you understand, uh, your user, your customer, um, uh, very deeply and actually bring AI to them, um, that's the category that I'm excited about. I want, I want, I want to see more proactive, uh, proactive AI products coming, coming into the market next year.

• 41:18Apoorv Agrawal?

  Yeah.

• 41:19Speaker 01

  That, that's one, that's, that is what is going to actually take it from a, uh, five percent of the users being power users to hundred percent.

• 41:27Apoorv Agrawal?

  Yeah, yeah, yeah. Your favorite AI tool that you use in your lives?

• 41:33Speaker 00

  I think Glean is awesome. I mean, if that was not clear. Let's go. So I use it all the time. I actually, a lot of the questions I would ask from the team, the thing you said you changed. I first ask Glean and then see, you know, if it nails it or not, then if it doesn't, then I'll spin up a 30-person team to go. And have three meetings and all that to get, you know, the explanation of some simple concept for me, but usually Glean nails it.

• 41:56Apoorv Agrawal?

  Yeah.

• 41:58Speaker 01

  For me, I'm excited about note takers. I've used Granola myself and Fathom and a few others. But note taking is actually fascinating. I mean, I think the, I feel like, you know, if you, if you take those notes and then if you utilize it the right way, like for example, what Ali was saying, like, you know, that becomes a source of what then actually creates knowledge, saves data in your systems, that's going to change, you know, how companies work.

• 42:24Apoorv Agrawal?

  Yeah. You know, in closing, I'd love to get your vision for your companies. We'll start with Ali's favorite tool, Glean. Congrats, you just announced crossing a big milestone, $200 million in revenue run rate. You're signing big deals, $10 million deals, you've got super users, I'm seeing, you're seeing casual users. Paint us a vision for Glean from here to a billion in revenue.

• 42:47Speaker 01

  I think they're still doing annual planning, which also some, you know, AI companies are telling me that's so old school. But we're doing it regardless.

• 42:54Speaker 00

  That's just because there's early startups. Like, did you do annual planning when you started Glean?

• 42:59Speaker 01

  No.

• 43:00Speaker 00

  No.

• 43:01Speaker 01

  So, but I think for us, the thing that I'm most excited about again is, so we think a lot about AI literacy and how do you get everybody along on this journey? And we're not seeing it right now. Like, Glean is a heavily used product, but, but still, there's a big variance between the top users and the ones, you know, at the bottom. And and that's what we want to change. Uh so for the future for us is we want to be we want Glean to be this um very personal companion. Um for every person in every company in the world. Um this this companion with which you know is is is you know you have a very confidential relationship with this companion.

• 43:46Speaker 01

  In the sense that whatever you ask this companion, you know, whatever communication you have with them. Um you know, it's it's fully privileged. Nobody else gets to see it. But this companion knows everything about you and your work life. It knows your day. It knows your week. Uh it knows who I even meet. Um um you know, in the day to day, it knows your weekly goals. It knows what you know, what things you're not good at or what your career ambitions are. And with all of that, you know, uh this this personal companion is um is sort of helping you now with your work. Um it you know, hopefully takes majority of your tasks automatically. Um um what what you know, works on them before you ask it to work on them.

• 44:27Speaker 01

  Uh and and that's so that's sort of the vision that we are, you know, taking our product to. We we have most of the found, you know, foundation for this in place already. Um today you have to come to Glean to get most of that work done. In the future, we want Glean to actually come to you and do that work.

• 44:40Apoorv Agrawal?

  Fascinating. Well, we can keep going for a bit, but I'm being called on time. Thank you so much for chopping it up with us. You got a lot of help, a lot of insights here. Really appreciate it.

• 44:49Speaker 01

  Thank you.

• 44:49Apoorv Agrawal?

  Thank you. Thank you.

• 44:53Speaker 01

  Thank you.

• 44:56Apoorv Agrawal?

  All right, gentlemen. Thank you so much.