Pioneer 2026 | Fin Fireside: Leading through the AI Transformation

Hey, thanks for having me. Appreciate it.

Great.

Sounds like history.

Okay. Okay.

Yeah. Micro OG level.

Yeah. Yeah.

Yeah. Right.

We had a couple recessions in there.

I think this is qualitatively very different probably from every one of those. Everyone's always trying to find the this is the next mobile, this is the next, you know, kind of SaaS, this is the next internet, and I've been guilty of doing that as well to try and analogize it. And then you're just like, you know, this is such a different scale of transformation because of the inevitable ripple in what the future of kind of industry looks like. And prior to AI, we had mostly technology transitions: mainframe to the mini computer, mini computer to the PC, PC client-server, cloud, SaaS, mobile. And at each one of these stages, work got more efficient, more of the world got interconnected. It did change fundamentally the workflows of a lot of these processes, but they were technology transitions. And AI is a technology transition, but also effectively a labor and work transformation.

And the easiest way to think about this is, for what we were building when we started Box, at the time it was a much simpler, faster, easier way to share and work with your enterprise data. But there was a version of that six months prior that worked just in your network, in your data center, on-premises, and we just made the faster, cheaper, lighter, easier version. AI doesn't have that. There's not like, well, you know, AI is just coding but faster and easier and more democratized. Like, no, no, it's a fundamentally reshaping of how you build software. It's a fundamental reshaping of how we deliver customer support and service. It's a fundamental reshaping to what a lawyer is doing with their time and their energy and how they work with their clients. So I think this is easily the biggest technology-driven revolution we have in an enterprise or in a work sense, at least certainly in our lifetimes. And so I think some of the analogies are breaking down, but at a minimum you can at least do the baseline of just like, you know, pre-PC, pre-internet, post-internet, it's that scale times, you know, 10.

Yeah.

But actually, but with a weird threading, because there's a most, you know, I don't know how much people are on like, you know, X or whatnot. Usually when people say this time is different, that generally puts you in a certain quadrant of the more risk or, you know, in the tech industry the word doomer sort of has emerged. And so if you hear this time is different, it's usually about, you know, more of the dangers and the risks of this. I'm a this time is different on how big of a technology change this is, but more I land in the positive side of the quadrant of now the upside of what that creates.

Yeah, I mean, mostly the things that I try and think through in our organization, we're about 3,000 people. We look like a classic software company in terms of where the distribution of work kind of tends to coalesce around in go-to-market and engineering and kind of back-office operations. I first and foremost look at a lot of work and try and imagine if that particular job had access to like 10 times or 100 times more information at its disposal, what would you do differently in that job function? So if you were a sales rep and you had literally 10 or 100 times more intelligence about your customer, your accounts you're going after, what they're doing, what they're interested in, what signal is happening in our product, how much better would you be able to go and actually deliver the right message, the right value proposition to that customer. And so then what's the gap between how we're selling today and a world where you had 10 or 100 times better information at any time? So that's sort of one vector, and I look across roughly most of the organization and say what could that do.

And then there's another vector, which is what parts of jobs are we spending an inordinate amount of time moving things between systems? Like we have to go from the ERP system to the HR system, we have to do this workflow where we're moving this type of data, and when are we at a point where agents are good enough at delivering those tasks in an incredibly high-quality way? And two years ago I would say it would be a waste of time to try and pursue any of that kind of automation. It just wasn't worth it. Like you guys know this pretty deeply. Like two years ago in the customer service agent experience, it was sort of like you could basically get an answer from a knowledge base and that was kind of like state-of-the-art, is like getting accurate answers from a knowledge base. Now in 2026, it's not only accurate answers in a knowledge base, but probably performing tasks, and probably actually those tasks are over time getting better and better because there's some kind of feedback loop and eval process and continuous improvement that's going on that was not architecturally possible even six or nine months ago. Like the models were just not good enough at doing that.

So now I look across the organization and I say, okay, where are there opportunities where with a good feedback loop, with the level of capability that models have at this point, where could we apply automation to get just way more throughput in that particular type of task, in that kind of workflow? And then the third thing is almost orthogonal to AI, which is, okay, if I've done these two things, we're getting way more information to sellers and they're able to be more productive, we're automating tasks that we just don't really want people doing. If you're doing things right, hopefully now you have a surplus. And then you're like, okay, well, what do I want to now go do that AI can't automate or hasn't at least attempted to automate yet? And now how do I go apply resources into those areas that previously I just couldn't fund, I couldn't invest in, and now I want to go actually supply resources into those areas.

So it's sort of like, what can you automate? Do that to the maximum degree, but then what do you now benefit from by having those extra surplus resources and how do we go apply those? And that's like the full loop, which is like, okay, we might automate one part of customer service just to go fund a different part of customer success. And so that just becomes this very positive flywheel that we'd love to run that flywheel as many times as possible.

Yeah.

They're all doing AI things.

Yeah. Yeah. So like kind of watch your.

Yeah.

Yeah. Yeah. Yeah. So, I don't know if it's realistic. I'll tell you my optimistic take based on the sort of setup. So, it was funny. About 12 years ago, maybe 10 years ago, I thought for some reason it would be a fun project to start a class at Stanford on.

Okay. Yeah. On business model disruption. And what was happening was this was the wave of digital disruption that was playing out, and basically the idea was like every single industry is going to be completely flipped upside down. You know, banking, travel, hospitality, transportation, everything. And we were actually right on the core theories, like basically the consumer patterns changed. Many of the business models of these industries kind of ended up evolving. But one of the core premise, basically the core premise of the thing was how would the incumbent survive with all of this startup disruption? And the part that we still got right, but it just happened at an even greater magnitude than I even would have predicted, is actually the incumbents ended up taking the digital technology and transforming their own businesses and services. And so then left with a choice for a consumer: do you want the incumbent? Like now consumers don't really think about incumbents, and that's like a tech industry invention. Like to a consumer, you're just like, do you want to keep banking at the place that reliably has never lost you money and has never gone out of business and has never had a bank run, or do you want the tiny little startup that's trying to disrupt that thing and we don't really know if it's going to survive? And basically by volume, most people still chose the incumbent, and it just that the incumbent was able to modernize because they had enough time. And it's this classic thing: can the incumbent go digital before the disruptor becomes big enough to matter?

And in that sort of period of time, the incumbents actually did quite well. Like JP Morgan is four times larger than when we started this class. Walmart is probably eight times larger than when we started this class. These were not the things that we would have expected to play out. So now do the AI version, and you're like, okay, so where is AI an advantage to a disruptor and where is it an advantage to an incumbent? And you can kind of go through and start to roughly get a sense of what might play out. So to a startup, it's a clear advantage in coding because now I can build so much more software way faster than the incumbent because the incumbent has to go and figure out its legacy processes and it's going to take them a while. So then you're like, okay, well, which industries is the speed of software delivery the big X factor in that industry's competitiveness? And it's like, well, lo and behold, it's the software industry. So we're seeing a dynamic around people saying, okay, well, do I want to swap out this vendor for something newer, but then at the same time, you have incumbents that are very much alive and at the wheel running things forward, Salesforce in particular being one of those, which is like, no, we're going to just transform the company. And we're going to make sure that we actually disrupt ourselves.

So, okay, so then you have the software thing. You have other examples of like, well, maybe I could use AI to dramatically accelerate life sciences discovery. And you look at that, you're like, well, still at the end of the day, are you going to trust Eli Lilly when they say they have a breakthrough? Are you going to trust the random company that you've never heard of? And it's like, you know, tie goes to the incumbent in something as important as healthcare or life sciences. You could just go through the list and you can see that actually existing businesses are in actually a structurally sound position.

With a massive caveat or asterisk, which is like if you transform, because at some point the consumer will just not have patience, like they will do the transaction from their AI agent, their muse, their instinct, their dot, and if your product or service is not set up to be able to respond to that, then what will happen is you'll just lose that transaction. So I think because the diffusion of AI takes long enough, it's giving incumbents enough time to rework, to retool. And so I'm much more in the, again, for incumbents, optimistic camp. And then even for startups, I'm actually pretty optimistic. It's just that they're going to have to go after different things than what the traditional incumbents have gone after.

I think that this is why I land in like a weird category in the industry, which is like I think it's still directionally good for a lot of categories. That doesn't mean it's easy for everybody. There's going to be like, I'm still working, I don't know, 14 hours a day just to make sure we can turn the crank as a platform. And so it's like not easy, but like I like our odds of when I kind of think about where the future's going and our ability to kind of be built for it. And so I go through the list of other existing platforms or even non-software companies, and I think with the right wiring you can kind of change the course of the ship. And then again I look at non-software companies and I actually think there's a lot of upside to AI that is like the classic thing in Innovator's Dilemma, and everybody should go reread the book from Clayton Christensen, and the core idea is that it's usually a business model disruption that drives the innovator's dilemma.

And the incumbent can't respond to the business model disruption because the new disruptor looks cheaper or faster or something that will look less profitable to the incumbent. So then you have to just look at your organization or your business model or your product and you have to say what parts of this does AI cause an innovator's dilemma moment for us, versus which parts are actually it's all just upside. Like again, an Eli Lilly looks at AI and they just say we can now run 10 times the amount of research experiments, and all of the infrastructure we have for testing that, for being able to do clinical trials, for being able to work through the FDA, for being able to get drugs to market, all of that is our strategic structural advantage, and AI is just an accelerant to all of that. So you kind of have to figure out where are you on the is this a disruptive technology? Is it a sustaining kind of positive technology? And I actually think a lot of incumbents benefit because AI is just an accelerant to their core business.