The Exchanges

Q And do you think it's necessarily humans have a seat and agents have consumption, or would there be an argument for saying that, uh, agents in some way are not that dissimilar from, from humans, although they, they'll be doing a lot more with a lot more, uh, volume of data, and therefore there should be some kind of, like, seat-based pricing for agents?

A I think, I think this was, this is sort of a tougher category because, um, it all depends on the agentic use case. So, um, like I can totally see a world where we already have some customers playing around this idea of like, should agents have a box seat? Because, because why? Because they actually need to store data, uh, that gets retained and governed over a long period of time. And you want to be able to track it and manage it just like a person, but it's gotta be stateful. And so that kind of makes sense as like, we have to give it a name and a thing in our system to make that work. Do we charge the same as a regular end user seat? Probably not. Probably it's gotta be cheaper. Um, but then there's a lot of situations where the agent doesn't need an ongoing seat. They just need to be doing a lot of operations, in which case it's, it's probably just pure consumption. So I think it really depends on where does your software category land on? Is there a reason why you'd have an agent be stateful in that organization? Um, and, and, and kind of take on an identity and take on ongoing work. Uh, versus it's a thing that just every employee calls on demand. And that, that would probably determine, you know, what that business model looks like.

AI assessment note: “it all depends on the agentic use case.”

Q and then we did a pilot and that never really worked out. So now you're coming back to me like two or three years later and you say, no, no, no, no, no, no. Agent is the thing. And like this time, if you don't do it, you're going to, you're going to die. In the spectrum between skeptical and enthusiastic, where would you put the mood in large enterprises?

A I would say if I did like the broadest sample, I think the mood would veer statistically more optimistic than maybe the framing that I think maybe you landed on like a few extra cynical CIOs in that. I think because what's happening is the CIO and our main audience is the CIO. Uh, when we talk to the CIO, they know that their engineering teams are using cloud code and codex and cursor, and they're seeing the productivity gains come out of those teams. And they're like, yeah, like my teams are just building, you know, apps way faster. They're, they're being able to tackle IT projects much more quickly. Like we're, we're, we're doing security reviews faster. Like they're seeing the productivity gains in their, in their function. And I think they're often saying, well, how do I bring those same IT, you know, productivity gains To the non-IT parts of the organization, and they're having the business pull them and say, I want access to, to co-work. I want access to codex. I want access to these tools as well. So there's, there's actually a certain kind of sex appeal to, to these tools right now where the business is sort of demanding. I want to be on the agentic train because I'm seeing all these, these great use cases. And so I, I think the tone is actually remarkably optimistic and excited and positive as opposed to, you know, there, there's a sort of, You know, typical trough of …

AI assessment note: “I think the mood would veer statistically more optimistic”

Q IPOs and we've seen companies that are compounding faster than ever. Where do you think that leaves startups, including vertical startups? Where do you see the opportunities? Are we in a, in a world where everybody is ultimately either an open AI or an anthropic employee? Or, you know, service, uh, industry, uh, supporting them? Or is there room for lots of people to do, uh, lots of different things?

A I remain pretty, pretty confident and optimistic on the, the need for a kind of a bridge layer from the AI capability to the end user workflow. And, and some might sort of say that this gets kind of bitter lessened out, um, which is, which is, you know, oh, these things are wrappers on the model. And, and at some point there's a training run where it just like is the final training run That makes the, the renders the, the kind of vertical app or, or function specific, you know, app, uh, you know, not as useful. And, um, and I think that is a little bit too much of an accelerationist view of what people are doing with the tool, which is like, it's not just like what the model is spitting out or the model's ability to review information. It is how was the thing wired up into the business workflow? How did it get the context that it needed to be useful? I think if you're in a, in an, in an industry or a line of business There's a heavy amount of, of kind of integration with data sets, heavy amount of, of kind of bespoke workflows that that company does. That usually means that there's going to be a need for change management, implementation, ongoing support, ongoing expertise. And unless the labs build out literally the equivalent of hundreds or thousands of people for every single vertical and every single line of business, that means that there's actually a lot of opportunity in…

AI assessment note: “I remain pretty, pretty confident and optimistic on the, the need for a kind of a bridge layer”

Q Okay. Because your head's down executing or because you don't think it's a problem?

A I don't think it's a problem. Uh, I mean, probably if I'm like really selfish, like just means more shareholders for us. Um, like if you just on a continuum said, would you like there to be, you know, 10,000 public tech companies or 1000, like where would you want to be on that continuum? You know, you can, you can just instantly imagine the like liquidity dynamics that would emerge. Um, so, so I'm kind of neutral to like, I don't really care that much. Um, I think the, the interesting innovation, uh, that has emerged is, is, and this is, it's funny because people kind of, people kind of don't know exactly the, the root cause, like some, you know, maybe people think like, 20 years ago, we made it so hard to go public. I'm actually not in that camp. I don't think we made it too hard to go public. I think actually it's good to have a, of, of an, a heightened degree of scrutiny and, and regulatory pressure on being a public company. It, it, It is, is absolutely net positive for the average shareholder that we have all of these systems and, um, and governance controls in place. I think that's only a good thing. Um, but, but no matter how we got here, where we're at is we now have this new innovation, which is late stage capital that can basically keep you private for as far as we can tell, maybe forever, because you can, you can kind of outrun the problem of converting, um, these c…

AI assessment note: “I don't think it's a problem.”

Q Yeah. And we were talking about Snowflake a minute ago, and they did their, you know, Arctic model, and, you know, Databricks did, like, PJRX. Yeah. Did you ever consider building your own model, or is that just completely outside of the business?

A We considered it for 10 minutes, and, um, and we said, uh, no effing way, um, are we gonna get in this war? Um, uh, so it's, uh, you know, back to the Bill Joy thing, it's like, It's like, why would you want to, why would you even want the, the, the, the brain, you know, aneurysm of, of thinking about like, oh, do we have the, you know, all the latest researchers in our company and we're competing for that and we need to buy two billion dollars of compute for the next training run. Like that is just a war that you want to ride the tailwinds of that. Not, not playing that. Now, different question if we were Meta and Amazon and those guys. Um, but, but, but the moment you're not those, you have to like understand where you are in the ecosystem and then figure out, you know, how to ride that. So we even like, I, I, we, we debated for more than 10 minutes, fine tuning stuff. And even then we were like, no, cause like, you know, there's just model breakthroughs every day. Like what, why, why, why, why, why even have anything internally that could cause us to confuse ourselves That like our fine tune thing is like really important. You know, there, there's a lot of things that you can do where, where, um, and actually this was, this started to become an interview question that I had, uh, for people coming into the AI team. I was like, do you want to fine tune a model? And, and if the…

AI assessment note: “We considered it for 10 minutes, and, um, and we said, uh, no effing way”

Q Very, very fair. It does seem to be, uh, a topic, uh, that's top of mind though. And again, I can, it'll be that, that tension between like the tech ecosystem, Silicon Valley, where like token maxing is, is really a thing, whereas large enterprises, uh, are worried about, uh, costs. So what do you hear and, uh, what do you recommend people? Do your customers.

A I don't know if I have good recommendations that actually, um, but I would say the token, the, uh, when we go and talk to organizations right now about where they are with agents, uh, tokens, the cost of tokens and budgeting and budget planning and all of this probably is at least one third of the hottest button issues that relate to AI. Um, and it might even be like tied for number one, like half the time, because what they've seen is, is this move from, You know, everybody started calling it like we were doing subsidization as an industry. I don't, I don't really think about it like that. I would say that the costs were just low enough that these things were included, like cursor just included, you know, a lot of usage and maybe it was subsidization, but it was actually just like they could model that under their subscription fee, um, you know, in a, in a, in a fairly clean way. And then all of a sudden what happened was these agents just can do way more work. Their context windows are way larger. The cost of inference is way more because they have way more parameters and their capabilities way better. So, so we've just gone from, you know, like the, a pricing model of a chat bot or like type ahead functionality and get up co-pilot to a, to that pricing model no longer working when, when one, you know, coding agent could be consuming, you know, a thousand dollars of, of compu…

AI assessment note: “when we go and talk to organizations right now about where they are with agents”

Q It's just a, just a DLP problem at worst, right?

A Just a DLP problem. And honestly, like in many ways, not that different from somebody going to Google to say like, I want to go research this customer versus going to chat to BT and say, I want to reach us to kind of like, like almost nothing has changed about the security paradigm of that enterprise. So maybe the prompt could include a little bit more IP, but like, But like the work you were doing was not like, like that, the blast radius of that work was, was kind of quite contained. Conversely, I go to an agent and I happen to have access to the Salesforce MCP server. And, um, uh, and, and it's, it's actually incredible. And it's actually one of the reasons why I totally believe in headless software, but I can do like, I could do a lot of work with that and I could pull out a lot of data and I can ask a lot of very powerful questions. And a company is going to have to say, well, should every employee have the same level of access? And how should we make sure that we've cleaned up our access controls for that? And how do we tell people again, like what types of queries should they be doing that are going to have different kind of cost profiles? And now you have to do that for each of your software, you know, vendors and applications. And then you have to figure out, like, what is the new workflow on the other end of this? Do you really want employees prompting, you know, thei…

AI assessment note: “Just a DLP problem. And honestly, like in many ways, not that different”

Q All right, let's get into, um, all the, uh, awesome things that you guys have been building at Box over the last couple of years. Uh, maybe give us a tour of, uh, from hubs to agents. Uh, what are the different moving parts?

A Yeah. So, um, so the way to think about Box is we're a platform that helps you store, share, manage, collaborate on data. We have a layer of, you know, security. We have a layer of file permissions. We have a layer of data governance. So that's what we've been working on for, you know, basically two decades. What we added was a layer, uh, which is just, we call it the AI platform. Uh, and it has kind of all the plumbing that you would expect you would need to do to be able to work on content in an AI. Uh, in an AI, uh, uh, context. And so, uh, so what do you need to be able to do? Well, you need to be able to like, you know, uh, some of it, some of it, by the way, benefits directly from other use cases that we've had. So for instance, for 15 years, you click on a document, you see it in your browser. Well, why is that possible? It's because we have a conversion engine, takes the Word doc, makes it into a PDF. We stream the PDF to the browser with, with PDF JS, and, and that's how you get it. Well, guess what? In that process, you've extracted, we've, we've, we, we can easily do text extraction. We can then do the embeddings on that text extraction. So that's a service that, that largely was built out for a different reason that now we kind of have as a standing start. We then put that into a vector database. Um, uh, so now we do, you know, we have, we have a layer that does the…

AI assessment note: “What we added was a layer, uh, which is just, we call it the AI platform.”

Q What is not working yet with agents? You know, a lot of people talk about how, uh, if you have chain of agents, then you end up with compounding errors. Have you experimented with that?

A Yeah, we we've seen every problem that that anybody's ever heard about. Um, so, uh, uh, You know, the more, the more agents you have, obviously you're like adding probabilistic on probabilistic on probabilistic. So like, good luck, like, you know, with, with what that could lead to, um, uh, you know, general search problems still exist, uh, at the end of the day, a lot of AI will be dependent on search technology and, uh, the quality of the search index, the quality of the ranking, uh, you know, most of our biggest ideas and biggest problems intersect with search. Um, and so, and so, you know, the moment that the AI finds the wrong thing, you know, the entire, your path dependent now on the wrong thing. So, so it's going to just make a whole series of, of bad decisions because it found the wrong thing as the, as the starting point. And we can just see this in our, in our personal lives, you know, using AI. It was funny. We asked a, um, we just do this, we, you know, a funny test. We said to every AI, uh, uh, kind of search product, um, uh, for consumers, we said, if you just do a query as simple as, as simple as, Tell me the last time, the last five times that the giants beat the Astros on a Tuesday. Okay. So you've got like, like, you know, a very complex problem in terms of the data set that has to be kind of combed through and then looked at from a logic standpoint and all o…

AI assessment note: “Yeah, we we've seen every problem that that anybody's ever heard about.”

Q I'm curious, um, as a public company CEO, um, Uh, not just, uh, you know, tariffs, but, um, how do you navigate uncertainty and volatility in public markets? You've, you've, you've been a public company CEO for 10 years now. Yeah. Do you worry about, uh, the ups and downs of the stock price, or do you have such a long-term view that it matters less?

A Well, uh, depends on if I can decouple the stock price from a more technical, uh, uh, you know, actual issue that, that we're dealing with. And so, I would say, I'm less worried about the stock price in the sense of moving because of, of Wall Street trading up or down and any particular, uh, you know, for any kind of particular reason, but, uh, very worried about, about the embedded issue that, that might be related to that. So, you know, let's say sometimes it's self-caused. So we have to execute better in a particular area. And the stock price in those situations is just, uh, another, uh, it's a symptom of the, of the actual thing that you're dealing with. And, and it's your kind of like, you know, real time KPI of, of that problem. And then in the case of tariffs, the thing that, that, you know, scares the living, you know, heck out of me is, is just actually the economic impact that this would, this would have. So I don't really care about the stock price, you know, it per se, I care about the health of, of the country economically. And then obviously there's lots of follow through that, that would happen and ripple effect that would happen if, if as of, you know, Tuesday, Uh, afternoon, you know, the, the tariffs that have been proposed, if those roll through, um, uh, you know, this is going to be a total disaster. So hopefully, whenever this airs, we'll be able to look ba…

AI assessment note: “I'm less worried about the stock price in the sense of moving because of”

Q Over the last, uh, couple of days, uh, Toby from Shopify had this very interesting tweet that I think you commented on, uh, about how he's, uh, effectively making AI usage, uh, mandatory within Shopify. I, I, I, is that something that you do that, uh, how do you, I guess, create an AI culture, uh, within Box?

A Yeah, so, uh, we, we have told everybody in the company that we wanna use AI, uh, to be as Productive as possible and aggressively use it, um, uh, across the business, you know, with an asterisk on, on certain kind of production use cases and whatnot. Um, and, uh, we, we do internal things where we, we have, uh, members of the team, uh, show up at our, our kind of internal all hands event, um, which is a weekly kind of, uh, a video call show what they're using AI for within box, uh, how it's helping them. So we're trying to get everybody's kind of just Just, you know, kind of creative juices flowing. I thought Toby's memo was fantastic. Um, it was, it, it, it hit on all of the, the, basically the, the kind of the, the, the core topics you need to start to think about as a company. Here's how we should start to experiment. Here's how we should, uh, you know, do product development. Here's how you should be thinking about, you know, when can this augment, you know, how we work and move faster. Um, so, uh, so I thought it was a, a great conversation starter for, for a lot of folks.

AI assessment note: “we have told everybody in the company that we wanna use AI, uh, to be as Productive as possible”

Q And you think that's, uh, mostly the chat GPT moment of like people seeing this in their, in their lives? Yeah.

A Without chat GPT, none of this would have happened. Um, now maybe, you know, a different timeline, somebody else would have done chat GPT and Gemini would have launched first and whatnot. So we, we could still have been on this timeline. I think somebody would have figured out like, Hey, why don't we put what, you know, GPT three into a chat wrapper, but like it could not have happened in a better sequence that it just exploded. Consumers got it. The next generation workforce Is like fully addicted to this stuff. So what, what, you know, you have, you actually have two really interesting pressures that, that are, I don't know if they're totally unusual, but, uh, but they, but this didn't happen because in cloud, maybe it happened in mobile, maybe it happened in PCs, but what happened, maybe it happened in internet actually. Um, uh, I just wasn't in a corporate environment to kind of feel it, but, but you have like the CEO is sort of like, Hey, what are we doing about AI? And then you have the new workforce coming in and they're like, I don't know how to work. Like if I don't have like a thing that can like help me answer questions and autocomplete stuff and do this, you know, like, I can't believe you guys type all these words out. Like, what are you doing with your time? Like, like it will actually, you know, it's, it's, I'm going through a journey mentally of this whole thing…

AI assessment note: “Without chat GPT, none of this would have happened.”

Q Uh, what's, uh, what are you guys doing in, in, in some granular detail, uh, around agents? What's, uh, what are you building? What's working? What's not working yet?

A Yeah. So, um, so, so back to the whole kind of layer of the, of, of, of how we think about an agent internally so far is, and these are the kind of core primitives, um, you've got access to data, you've got the user permissions, you have a set of tools in box. The tools right now are things like, um, uh, you know, crack open a document and, you know, feed the model, um, you know, uh, uh, you know, chunks, uh, based on the user's query, Uh, search is a, is a tool we're working on. So you have a set of tools, and then you have the model. And then we're adding, you know, kind of an agentic workflow layer as well. And in combination, you can just think about it as, you know, agents for content. And so whatever you would, whatever you think a human does with, with documents and content, we should have an agent layer that lets you build, build that exact thing. And so what's amazing for us, and this is where I get most excited about AI, is In our business, I'm going to make up all the stats. None of the numbers are exactly real, but I would argue that probably for 95% of our data, people have things that they would love to be able to do with that data that they just don't ever do because it's too expensive. It's too time consuming. They didn't think to do it. They don't have a person with that expertise to do it. And so, and so basically it kind of means that like 95% of your data is…

AI assessment note: “you've got access to data, you've got the user permissions, you have a set of tools”

Q Fascinating. What do you recommend that people do in your conversations, given this litany of Things that need to happen and the space of innovation, all of it happen at the same time. You mentioned internal FDs. That's super interesting. We can talk about external FDs, which I think is a better understood thing. Where should people start or how do they accelerate?

A So the one part where I'm just like, I, you know, I'm a, I'm a hammer looking for nails is I see most things as a data problem, um, uh, and, and, and data with associated things like access controls and like how well defined is the workflow, et cetera. So most agentic challenges, I, I think are kind of inversions of, of, of basically like you have a data challenge, like the agent can't get access to the right information to do the work. Uh, maybe they have access to too much information, in which case then, then they're just going to like roam around and do the wrong thing, or they have access to too little information, in which case obviously they're not going to work, or they don't have enough context to be able to execute the task. And which means they need more information You know, surrounding the task. So, so we, we see data problems everywhere, um, that, that we look. And so I think one of the first steps is like your enterprise just needs to be prepared from a, from a data standpoint and from a, from a kind of a core architecture. And I think we, for 20 to 30 years in IT, it was sort of okay to, to sort of have all these systems, some redundant, some not well managed. You could kind of throw humans at the problem and just sort of say,

AI assessment note: “one of the first steps is like your enterprise just needs to be prepared”