3546: Box and the Leadership Shifts Behind Becoming an AI First Company

[00:00:04] - [Speaker 0]
Welcome back to the Tech Talks Daily podcast. I've got an exciting guest with you today. I've got Olivia joining me from Box. She's one of the most experienced operators in enterprise technology and go to market leadership. She has led global teams at Google, Google Cloud, Dropbox, Notion, and now Box, where she oversees marketing, sales, customer success, and partnerships around the world.

[00:00:32] - [Speaker 0]
But predictably, yes, we're gonna be talking about AI, how it's moved from experimentation to execution, and my guess brings a grounded view on what it takes to become an AI first organization. And I think we're gonna hear that phrase more and more throughout 2026. Few years ago, it was every business is a tech business. Now I think we're gonna start I think we're gonna be hearing more and more about an AI first organization and how productivity gains come from redesigning how work happens and not simply layering automation on top of old processes. And, yeah, you know who you are.

[00:01:11] - [Speaker 0]
So I'll be very interested in your thoughts, your takeaways on this, but, I'll check-in with you all at the end. But before I get my guest on today, I wanna give a quick thank you to my friends at Denodo. Because one of the questions I hear more and more from listeners on this podcast is, why does AI succeed or why does it fail? Because let's be honest, AI is moving fast, but success is often still elusive. Now most projects fail not because of the AI, but because the data foundation isn't ready.

[00:01:46] - [Speaker 0]
So if you're ready to understand why your AI projects fail and how to succeed with AI, simply visit denodo.com and take control of your data world. But now I would like to invite you and sitting down with myself and Olivia right now. So a massive massive warm welcome to the show, Olivia. Can you tell everyone listening a little about who you are and what you do?

[00:02:11] - [Speaker 1]
Well, first of all, thank you for having me, Neil. It's a treat to be on your show. And diving right in, I am a COO at Box, which means I'm responsible for everything facing customers. So marketing and sales and support and partnerships, operations, all the ways in which we get to engage and and just drive impact with our customers, which is fantastic and keeps me going in the time of AI. It's been it's been quite a journey to take our customers on, so super excited to be here to talk about it.

[00:02:42] - [Speaker 0]
And I'm glad you join me. And doing a little research on you, you've been on a journey yourself, almost tech royalty, may I add. But you had senior roles at Google, Dropbox, Notion, and now, obviously, Box. I've got to ask from you're looking back at your career, how experiences shaped the way you think about AI as a a driver of change across the enterprise? Because you've probably seen this grow so much and evolve so much long before ChatGPT even arrived.

[00:03:10] - [Speaker 0]
You know?

[00:03:11] - [Speaker 1]
Yes. It has been honestly quite a journey. And even when you're sitting in the heart of Silicon Valley where it's all happening around you, even then you feel like, wait. Did I get all the facets and all the dimensions of the transformation that really needs to occur so it keeps you on your toes? At Google, we we obviously, it's like the the home of of Gemini, and everything that comes out of Google from an innovation perspective is amazing.

[00:03:36] - [Speaker 1]
But, also, Box works very closely with many of the other providers like Anthropic. And and so we and OpenAI. And so we really are kind of in the mix here. I would say from a journey perspective, it's really about the transformation. It's the AI itself is super amazing.

[00:03:59] - [Speaker 1]
The functionality is incredible. You know, labs here at Stanford were the ones really pushing on LLMs and made some of those amazing, amazing breakthroughs. But I would say it's the intersection now of generative AI with content that allows us to tap was basically dark data, dark unstructured data. Right? And so that is a unique time.

[00:04:23] - [Speaker 1]
That really is a unique time for us to be in an industry where we can help companies, nonprofits, communities truly transform the way they do work because they're able to tap into this unstructured data.

[00:04:39] - [Speaker 0]
And there are many, many reasons I was excited to get you on the podcast today, especially after the latest state of AI in the enterprise report landed on my desk. Set off a few text text by the sensors, and the report shows that early adopters, they're already seeing productivity lifts. They're close to 37% as well, which is a big number. But I'm curious. We've been talking about this for a couple of years now, but what separates those companies from those that are still stuck in experimentation and can't get out of pilot phase or or a little too overcautious?

[00:05:13] - [Speaker 0]
What is that is that gap getting wider now?

[00:05:17] - [Speaker 1]
Honestly, I do think it is, and I think it comes back to having a clear strategy on AI. And, of course, there's the old saying of execution needs strategy for lunch. So you have to pair. And as having been, you know, a strategy consultant for for many years, I'm very acutely aware of that since I then went into the operating role and saw how hard execution is. But, really, what it is is you have to have a clear idea of what you're trying to get done with AI.

[00:05:46] - [Speaker 1]
Like, yes, there is this concept of let a thousand flowers bloom, let everyone experiment, and that is fantastic and wonderful for people to understand what's the art of the possible, right, and to invent and create. But at some point, right, pretty early on, ideally, there has to be some sense of, okay. Where are the main places where we're gonna drive true, true transformational impact with AI? And let's lean in there. And let's make sure that the agents that we're deploying are instructed well.

[00:06:18] - [Speaker 1]
They continue to be coached. They have access and are connected to the right curated data. Right? And then that we actually drive the change management of adoption. And that's where it's all the really hard part because that's human behavior.

[00:06:32] - [Speaker 1]
Right? And that's teams adopting AI. And so that's the part where I see some companies really separate themselves from the ones who just kinda continue to dabble in it.

[00:06:44] - [Speaker 0]
And when I was doing a little research on you before you joined me today, I was reading how you often talk about that shift from simple automation to full AI driven redesign of of how work actually gets done. So when you look across your customer base there, where do you see that a shift like that gaining real traction? Any stores you can share? I don't have to name any names, but I'm just curious what you see there.

[00:07:07] - [Speaker 1]
We do. I mean, honestly, if you kind of think about the arc of how people get work done, right, and not to date myself, but if you go back to a time where outsourcing was a thing. Right? And and there's this epiphany that, oh, maybe we can drive efficiencies by outsourcing tasks. There was this key key learning that came out of that phase of of business that was, hey.

[00:07:31] - [Speaker 1]
If you outsource a convoluted workflow, you're probably no better off. So you gotta kinda clean it up before you outsource it. And I think that learning now applies kind of in a next gen version, but with AI. Right? Which is you actually can't just send the same process and and insert AI agents or insert AI functionality into it.

[00:07:55] - [Speaker 1]
You actually wanna step back and rethink the process knowing what is possible with AI. Right? And and I think about that even in the way we get work done at Box. Right? We have historically, we have flows of, people doing handoffs in order to kind of aggregate information in order to show up to the customer in an effective way.

[00:08:19] - [Speaker 1]
Right? But now, actually, you can have, our account executive cover all that ground with an agent. They don't actually need to wait for handoffs from three other teams for all those different pieces of information. They don't need to wait for the competitive intelligence to come out of the competitive intelligence team, the, you know, market context to come out of the product marketing team, the, you know, proof points to come out of the product team. Those agents are just dipping into those knowledge bases and aggregating all that information and just serving it up to the account executive.

[00:08:54] - [Speaker 1]
Right? Now it is the responsibility of those other teams to curate the information to make sure that when agents tap into it, they're getting insights and valuable information out. Right? So it's almost a new role for them, but the workflow itself has changed dramatically.

[00:09:12] - [Speaker 0]
And I think it was in November 2024 where Gartner famously said, AgenTik AI will dominate 2025. Predictably, it did, or at least it did in a lot of conversations. I went to that 25 tech conferences, AgenTik AI was mentioned in all of them. But I think 2026, this is where AI agents are now starting to take on meaningful tasks across systems. We're coming out of pilot.

[00:09:34] - [Speaker 0]
These things are going out there now. So how should leaders be thinking about adding these agents into the workforce, but doing so in a way that improves outcomes without creating new forms of risk? And we've kinda seen what happens there when you move fast and break things without thinking of the consequences. So what should leaders be thinking about this year, do you

[00:09:54] - [Speaker 1]
Absolutely. Very much generative AI was in the zeitgeist in 2025. I would say it's actually gonna be in our working models in 2026. Yeah. And it goes back to this key elements of how to be an AI first company.

[00:10:10] - [Speaker 1]
The first is that you really need to have an understanding of what you're trying to get done leveraging AI and to treat it as a source of growth. Right? If if you come into it with just purely an efficiency mindset, I think those teams are gonna basically do themselves a disservice for all of the upside that that's possible. I would say the second is that companies really need to be treating unstructured data as an asset. Right?

[00:10:39] - [Speaker 1]
So as they go into 2026, they need to actually have a content strategy, and they have to understand deeply where their most important unstructured data sits, right, and how they're gonna leverage that information with agents. And third, I would say it's all about governance. Right? Because of exactly what you're saying, there is risk if you put agents to work on data that is not well managed, that doesn't have the right permissioning, that doesn't have the right security protocol. Because as we know, these agents are pretty endeavors.

[00:11:12] - [Speaker 1]
They are agentic. You're asking them to do a task, and they will find all sorts of creative ways to either, invent data that they can't find or just retrieve any data regardless of whether it's, you know, got the right time stamp from the relevant person or whether, frankly, they should or should not have had access to it unless they're really controlled in a very serious way. So I would say that governance element, I think, will really, really come to light here in 2026 as people start actually putting agents to work.

[00:11:45] - [Speaker 0]
And I think around twelve months ago, many businesses that jumped on the AI bandwagon without a strategy were starting to report that, hey. We didn't know what problem we were solving. We're struggling to get an ROI from this project, and a lot of lessons we've learned. And and things have changed, since then. And refreshingly, your research shows that companies using a mix of models and agents tend to see a much higher ROI.

[00:12:09] - [Speaker 0]
So why is variety so important, and how do you help customers choose that that right mix for their environment? Because I would imagine it would differ between business or industry. Yes.

[00:12:21] - [Speaker 1]
What we see is that and not surprisingly, different LLMs perform different tasks better than others. And so and and we know that because we're working with content and unstructured data all day long. And so what we take is a very Switzerland approach to which models our customers want to use. We really offer all of them up available to them, and it's really a question of what the customer wants to do. Now we do default to the one that we know or we believe works best on the different tasks.

[00:12:52] - [Speaker 1]
Right? So we haven't a very opinionated view on which is the recommended model, but we always give our customers that ability to move between models because they might have different reasons to do so. Right? But we definitely see that the better outcomes come from companies who open themselves up to leveraging different models. And the beauty of the Box platform or any other, company that offers kind of the different models is that, is that we provide the ability to choose.

[00:13:23] - [Speaker 1]
Right? Our customers don't have to go out and actually be doing contracts with all of these different LLM providers. They just are a customer of Box, and then they get access to that level of choice and agency, and that's really important. Right? So and we also guarantee the the governance, the permissioning, all of the security that comes with it, and that's important as well.

[00:13:45] - [Speaker 0]
I just wanna give a big thank you to my sponsor who is supporting every show, every episode across the Tech Talks network this month. And this month, I'm proud to be partnering with Alcor. And anyone who's tried to scale an engineering team across borders, they will know firsthand how messy it can get because they deal with endless providers, then there's confusing rules to deal with in each and every region and fees that always seem to surface at the last minute. Now, Elcore, they solve that by acting as a partner rather than just an intermediary, And they focus on tech teams that expand in Eastern Europe and Latin America, and they bring employer of record services together with recruiting. So, essentially, they help you pick the right country, source the right engineers, and assess them properly, and then get them active for you and your company within days.

[00:14:38] - [Speaker 0]
And one of the things that stands out for me is the financial transparency. Around 85% of what you pay goes directly to your engineers. Their fee goes down as your team grows, and if you ever wanted to bring your team in house, you do so with no exit costs. That kind of clarity is why Silicon Valley startups, including several unicorns, have chosen Alcor, and you can find out more by simply going to alcor.com/podcast or follow the link in the show notes below. And governance is becoming a core factor for evaluating AI platforms.

[00:15:14] - [Speaker 0]
So I'm curious. Where where do you see the biggest blind spots in how many organizations are approaching AI governance today? Because there is a lot of talk around it, but there are blind spots. And and indeed, how are you at Box helping to close some of these gaps too?

[00:15:28] - [Speaker 1]
Yeah. I mean, I would say that, you know, legacy interfaces and, older constructs don't actually work well in this new agentic AI environment because we have to assume that these agents are acting autonomously. Right? They're as I said, they're quite endeavors. And so we really need to underscore from the beginning a a governance structure that contemplates, these autonomous agents.

[00:16:00] - [Speaker 1]
Right? And then put in the security and the governance and the permissioning in a way that anticipates that. Right? And so that's really important. And it will become even more complex and even more important as agents start talking to agents.

[00:16:15] - [Speaker 1]
Right? And especially when agents from different platforms start talking to agents. And we see that even in our customers today where they say, well, we trust the box agent, but we don't know if we trust all these other agents. Right? And that's an important understanding from an architecture perspective of, yeah, how are you designing the way in which information is exchanged?

[00:16:37] - [Speaker 1]
Right? And how do you make sure that you anchor your strategy in the agents and the AI that you do trust and that you do know is permissioned and governed, right, and that you make that the core? Right? And then the way in which information passes in and out is more governed by that. Right?

[00:16:54] - [Speaker 1]
And so you know it's more a dependable approach.

[00:16:57] - [Speaker 0]
And, again, over the last few years, we've gone through various stages from experimenting and playing into test, into production, and now we're going beyond all the hype that has filled our newsfeed. We're starting to see real maturity here. And one of the reasons that I wanted to bring that up is that your white paper outlines five stages of enterprise AI maturity. So which stages do you see most companies sitting in right now, and and what steps help them progress to something that that scales across their business?

[00:17:28] - [Speaker 1]
I would say that most companies are pretty early on on this journey. I mean, we describe these stages of almost being in the, okay. I'm using AI to get insights. Then you progress to we're using AI to do a task, and then you progress to we're using AI to do entire workflows. And then you progress to, okay.

[00:17:49] - [Speaker 1]
Actually, the AI is running the workflows, and humans are kinda checking in on it. Right? And then, of course, you kinda have this fifth stage, which is almost like the autonomous driving car. Right? Agents in the loop.

[00:18:01] - [Speaker 1]
That's kind of the key core principle that we have for AI first. But we do find that most people are kind of in that early early couple stages of, you know, they're using AI to do get insights, or they're using AI to do a task. And so what we're doing is we're working with our, with our customers to actually take them on that journey. I would say a number of our customers got a little burned early on. Right?

[00:18:28] - [Speaker 1]
They tried some of these more, I would say, like, personal productivity AI, tools that cost a lot of money, and they deployed it all because they had this urgency or they were getting pressure from their board to, quote, unquote, deploy AI. Right? And so they realized that, actually, unless you have a plan behind it and you understand how you're trying to impact your business, you're actually not gonna get a strong ROI. And, frankly, it will lead to a lot of frustrations and risky behaviors. Right?

[00:18:58] - [Speaker 1]
So what we try to do is meet the customer where they are, understand, of course, their security environment and everything they need to do to protect their data, and then take them on this journey of transformation.

[00:19:11] - [Speaker 0]
And with AI reshaping the competitive field, how are established companies using AI to to keep pace with and and sometimes outperform some of those younger AI native players. You mentioned there that a lot of businesses got burned by, going after the personal productivity side of things. So what what are they doing right, and how are they keeping pace with these these younger companies that are are running away with it at the moment?

[00:19:39] - [Speaker 1]
I would say that incumbent companies obviously have this amazing, amazing asset, which is their customer base. Yeah. Right? And so when they're leveraging AI, they're often leveraging AI across a vast surface area of customers that is just fundamentally different than a start up. Right?

[00:20:00] - [Speaker 1]
A start up has the benefit of being smaller and nimbler and maybe getting products out faster. But for distribution itself, they still have a big hill to climb. Right? And so when you think about other players that are more established like what you're describing, how they're thinking about AI or at least the ones that pull ahead, how those are thinking about AI is about innovation, is about bringing a better customer experience to their customers, about thinking about new and different ways that honestly were never possible. Right?

[00:20:31] - [Speaker 1]
You see a lot of the customization work that's happening in AI where companies are able to show up to the end customer with a wealth of understanding of everything that customer needs from from a a platform. Right? And so they're able to deliver impact in a way that I think is truly differentiated. And I would say that's true for us at Box. Right?

[00:20:53] - [Speaker 1]
We know our customers and have such a longitudinal surface area with our customers that then when we apply AI to all of the information about our customers, which, again, often sits in unstructured data, that we're then able to meet our customers in a way that's truly differentiated. And that part is really fun. Right? It's nice to see the customers, really engage with these tools and then turn around and transform their businesses.

[00:21:21] - [Speaker 0]
And one final step from your report is 90% of companies are planning to increase AI investment in the coming year. That tells me if you thought the pace of technological change in 2025 was fast, buckle up because it'll probably never go that slow again. Things are gonna continue to race ahead. And I've got to ask you at the beginning of a new year, what excites you here? What are you any trends you're watching?

[00:21:44] - [Speaker 0]
Anything just in general excite you, about the months ahead?

[00:21:50] - [Speaker 1]
Well, we're in this exact same process. Right? We're an AI first company, and we're looking at what we want to invest in next year. And we are making those same decisions to say, okay. We wanna invest in AI technology.

[00:22:03] - [Speaker 1]
It's our own technology. Right? So we run most of our agentic workflows on Box itself. Right? And we know that we wanna spend the time and actually coach the agents, instruct the agents, and make sure they're embedded in the workflows.

[00:22:17] - [Speaker 1]
So for us at Box, well, the way that shows up is actually the plan of running a a team that manages agents. And so that's been almost the mental flip. Right? And we have a team that comes together, which is our AI operating committee. And, basically, it has functional leaders from everyone in Box, and they come with the plan of what is what are the Box agents that they want to be deploying in this coming year to get their work done.

[00:22:50] - [Speaker 1]
And this is coming off the back of a quarter of pilots, right, where team members are testing out different agents, and then, a choice to say, okay. These are our two, three hero agents that we're gonna deploy in lead generation, in marketing, in engineering, in product. Right? And we all come together and we say, okay. These are our hero agents for each of these functions.

[00:23:14] - [Speaker 1]
Right? And our, teams are ultimately saying, okay. In 2026, these agents are actually gonna interact with each other. Right? So let's plan a world of that.

[00:23:24] - [Speaker 1]
And then let's also think about if you're a manager who manages agents, how are you gonna manage those agents? What are you gonna hold those agents responsible for? And what is a function are we expecting out of our agents who are working for these agent managers? So it's really like a whole flip in terms of how you think about a business, how you set expectations, how you set, really creative and innovative goals. Right?

[00:23:50] - [Speaker 1]
And so it's it's been a journey. I mean, I've been operating for so so many years. This is a whole new muscle. I turned to my leadership team the other day. Was like, we're gonna remember this.

[00:24:00] - [Speaker 1]
We are gonna remember 2026 as the year we had to figure out how to manage agents. And that's fun. I mean, if you think about that even from a career perspective, right, that's a I mean, hard, interesting, complex, but it's a really fun journey to be on. And so I'm just super excited about all we're gonna learn in 2026.

[00:24:21] - [Speaker 0]
Awesome. Exciting times ahead. Now for everybody listening, I'm gonna put, links to the state of AI in the enterprise report. I'll also post a link to, becoming an AI first company, the white paper there, and your LinkedIn. Is there any anywhere else you'd recommend people listening check out after listening to our conversation today?

[00:24:42] - [Speaker 1]
Yes. Well, absolutely. The becoming an AI first company is one that kind of we poured our heart and soul into, and and we have actually articles that follow that. So if you go to, our LinkedIn, you'll see reflective articles coming out of that, which is to say, here's what we set off to go do, and then here are all of our learnings along the way. And we're now on article three.

[00:25:05] - [Speaker 1]
So we're gonna keep on posting about what we're learning as being an AI first company, and you can find that on our LinkedIn or on our website. And then the state of AI and enterprise, yes, I would say is a is really something we we took away from all of the insights that our customers provided to us. Right? And so that was really wonderful to learn from everything they're going through and then be able to play it back to them and help them on that journey.

[00:25:31] - [Speaker 0]
Awesome. Well, I'll add links to everything, and I just love chatting with you. We covered so much in a short amount of time there from AI agents, how they're a game changer and rapidly joining the workforce, and AI, how it's disrupting the competitive landscape, but also how companies are boosting flexibility and ROI and multiple AI models and agents. So many big talking points around that. But at the heart of it all, we've also got data security and governance are key when evaluating AI platforms.

[00:26:01] - [Speaker 0]
As an ex IT guy, that makes me sleep well at night. But more than anything, thank you for joining me today and sharing your insights.

[00:26:08] - [Speaker 1]
Thank you, Neil, for having me. It was a treat.

[00:26:11] - [Speaker 0]
A big thank you to my guest for offering a clear eyed perspective on why AI first companies are beginning to pull ahead, but why governance and security and choice of models, how these things now matter just as much as speed. And whether it be agentic workflows or measurable productivity gains, I think today's conversation cut through some of the noise that typically surrounds enterprise AI adoption. So you'll find links in the show notes to Boxer's latest research, that state of the AI enterprise report, and the becoming an AI first company white paper. Take some time to go away, digest that, think about this conversation, and then come back to me. Tell me what resonated with you and whether your organization is still experimenting with AI or already seeing real returns.

[00:27:05] - [Speaker 0]
Pop by the techtalksnetwork.com website. You will find 4,000 interviews across nine podcasts. Sadly for you guys, they are all hosted by me. So if you're hoping for a different voice over there, you're have to hang on at the moment. Work in progress.

[00:27:23] - [Speaker 0]
But right now, you and I are in this together, and I will return again tomorrow with another guest. Thank you so much. Speak with you then. Bye for now.