Box, Inc. (BOX) Earnings Call Transcript & Summary

Hi. My name is Josh Baer, software analyst here at Morgan Stanley. We have the Box team. Before introductions, a brief disclosure. For important disclosures, please see the Morgan Stanley Research Disclosure website, www.morganstanley.com/researchdisclosures. If you have questions, reach out to your sales representatives. Thank you so much, Aaron Levie, CEO, Co-Founder; and Dylan Smith, Co-Founder and CFO of Box. Really excited to have you here today.

Thanks for having us.

Aaron, I want to start by asking you about your enthusiasm right now.

We're too excited? Very excited.

You've made comments like the current product lineup is the most transformational in the history of Box. And you have said that you are more excited now than ever. Can you help unpack that really what's making you so optimistic?

Yes. So we basically -- we are fully convinced that of all the software categories where AI will have a massive impact, we have landed on one that will be in the upper decile of transformational use cases. And the reason for that is what we help customers manage -- 115,000-plus customers, what we help them manage is all of this unstructured data, contracts, invoices, digital assets, blueprints, project plans, CAD files, bank statements, all of this data that is really only useful when a person looks at it. So you're looking at a contract and you're like, okay, these are the clauses in the contract that are risky. I need to review that. I'm looking at an invoice and I pull out the data from the invoice and put it into a database. I'm looking at a 20-page product requirements document, and I read it and I kind of understand it and move with the project forward. All of that data is incredibly valuable, but only when a person goes and touches it. AI basically lets us, for the first time ever, do the exact same value that a human does, but now we can digitize that value and then we can automate it. And so for the first time ever, we can start to really transform how we work with our unstructured data because humans can only look at so much data -- it takes a certain amount of time to look at this data. And in particular, what happens is you create something, you share it, you collaborate on it, but then it usually goes somewhere and you don't really touch it again. Now all of a sudden, all of that information has an incredible amount of value in it that you can finally tap into because the AI can go reach into that data, pull out what it needs and produce something of value. So for us, we kind of see it like a 10x increase in the value you can create from all of this unstructured information. And the sort of signal of this, we have financial metrics that are a signal of getting customers into Enterprise Plus, the very early traction of Enterprise Advanced. So those are very important signals. But qualitatively, the kind of CIO conversations we're having, global CTO conversations, we have Fortune 500 CEOs that are now starting to say, "Oh, am I actually getting enough value from my unstructured data? And what could AI do with my information that I didn't realize that it could do? I was not talking to Fortune 500 CEOs about file sharing and collaboration 10 years ago. That was like not an interesting topic for a global Fortune 500 CEO. It is a very interesting topic to say, are you actually tapping into the insights in all of this data that your company is producing? Do you know which customer you need to be leaning into? Do you know which part of your supply chain is slowing down? Do you know what new products you should build based on the customer data that you have? Those are all now use cases that Box can deliver and is frequently -- and increasingly delivering for customers. So it's a complete shift in what we can do. And so that would be exciting enough. But then you add to it, we have the right architecture, which is by -- from day 1, we are a multi-tenant SaaS cloud-first platform. We don't have a single customer on any other architecture. So there's a lot of companies that are like, oh, AI works for this version number of the platform you're on. For us, 100% of our customers have access to this technology if they upgrade into the relevant plan. Every single file in our platform, it automatically works for, barring certain file types that AI models can't support yet. And then we're moving at a pace and with a technology team that I think is letting us stay at the forefront of this industry in a really compelling way. I'll give you an example of that. Last week, OpenAI released GPT 4.5, and we were an early tester of the model. We were one of the only companies that put out an eval of enterprise -- of GPT 4.5. We sort of showed how it evaluated against GPT 4.0 for metadata extraction from documents, which is a pretty critical use case for enterprise workloads. We're the only company that put out an eval on day 1, an hour after the announcement from OpenAI, and that was because of our partnership that we have with OpenAI. Two days prior, we did the same thing, not with as much of our public eval data. We're still kind of working on that with Anthropic, but we were a day 1 early release partner for the Anthropic model, Claude 3.7. So we're at the right place, the right time, the right architecture and with a set of partnerships because of our kind of neutrality in the AI space that lets us now take advantage of all of the innovation that's happening. We work closely with Google. We work closely with OpenAI. We work closely with Anthropic. We intend to work closely with xAI. And you'll just see us continue to establish more and more of these types of partnerships. We have a great partnership with IBM, so you can imagine the kind of things that would happen with things like watsonx and the open source community. And so we just want to be the place where when you want to work with your unstructured data and content with AI, we want to be the best platform for that.

A lot to be excited about.

Yes.

You mentioned some KPIs in sort of tracking this progress. So I want to loop Dylan in. Where can we see this so far? And like where will we be able to see the momentum as the year goes on?

Yes. So ultimately, we'd expect the impact of everything around AI and Enterprise Advanced to show up in a lot of the metrics that we already disclosed around things like the net retention rates as just customers adopt this, certainly, the penetration of Suites overall as this both grows the footprint of existing customers as well as gets more and more of our customers to move into Suites, and those types of metrics. And then in the interim, I think as things are kind of scaling out, I think it will be more around here's some of the color commentary, overall magnitude of the impact in period and how things are evolving. But eventually, and as the model evolves, you could see us breaking out some of the components that more directly show the impact, whether that's on a quarterly or annual basis. And that's kind of categories like just the way that our consumption business and platform monetization is evolving to some of the more specific metrics around Enterprise Advanced, just as we gave Enterprise Plus metrics as it was ramping a few years ago until ultimately, Enterprise Plus became synonymous with Suites. But -- so I think that's kind of the types of categories you can expect to see. And then ultimately, some of those metrics are going to be around what we see as what's most meaningful to really understand the momentum that we're seeing.

Okay. Helpful. Want to stick on Suites, and there's a lot of early success around Enterprise Advanced. You called out several dozens of deals closed in Q4. It was only GA, I think, the last couple of weeks of the quarter.

Yes.

So what is really driving that momentum? Help us understand like what is it in Enterprise Advanced that's incremental to Enterprise Plus?

Yes. So it's actually the biggest step change that we've had in a new plan introduction in the company history. So we have now almost 2 decades of kind of learning how to introduce new plans. And usually, what would happen is we'd have sort of 1 to 2 breakout features that we identify and say, okay, that's sort of deserving of now a new plan type. 10, 15 years ago, it was single sign-on and some data governance functionality. And then with Enterprise Plus, it was our more advanced data security module, Box Shield and our data governance module we kind of pulled together as one suite. This is the biggest change based on the amount of timed -- kind of well-timed functionality we had on day 1. So the suite encompasses Box Apps, which is our no-code application building kind of foundation. So anything you want to do where you want a custom interface and a custom set of ways of working with your data. So let's say you want to be able to work with digital assets or invoices or a clinical drug trial process, you probably want an interface that's sort of designed for perusing, getting insights from that data, moving things through a process. So Box Apps lets you now do that. This is a plan that also has the first addition of our AI-based metadata extraction. So you want to be able to pull out the data from your documents. And so Box -- so Enterprise Advanced lets you do that. We also have new capabilities around being able to create forms to pull in data into the system and then a product called Doc Gen for being able to generate documents on demand, a statement, an invoice, a resume, a contract where you give us a template, and we'll go and produce that content for you and AI certainly enhances that over time. And then some data archival functionality. And then finally, the AI Studio. So if you want to be able to go create a custom agent for some particular workflow in your business, our AI Studio is available in Enterprise Advanced. And so it's sort of -- if you think about it symbolically, what it will represent is when you want to do AI-oriented workflows on content, Enterprise Advanced will be the plan that delivers that. And so over time, I think -- and this won't maybe be as sort of explicitly written out anywhere, but it will be how our sales motion works, which is our core plans, which eventually will kind of -- you'll feel like Enterprise Plus is kind of included in this are really for productivity, collaborating with your -- with people or with AI, being able to ask questions of AI, being able to summarize information, find what you're looking for, produce content using AI. And then Enterprise Advanced is when you want to go deploy AI agents to automate workflows, you create the financial analyst agent, and that's going to do some kind of long-running work and come back 10 minutes later, having completed a full research assignment, let's say, or an invoice automation agent that goes and looks through all your invoices, moves them through the right steps in the process, et cetera. So that's kind of where Enterprise Advanced will wear, where the core plans, including kind of Enterprise Plus will help you with more of that kind of end user kind of engagement with AI at a productivity level.

And so is the adoption path that we saw from Enterprise Plus, is that a good proxy for Enterprise Advanced? Obviously, there's a lot more capabilities from day 1, but thinking about just customer demand.

It's really tricky. This conference is sort of well timed to only have 2 weeks of data that's publicly available for us to be able to talk about. So 2 weeks in the grand scheme of Enterprise Plus and its sort of predecessor suites were -- have been available for 5.5 years. So I'd say maybe like let's check -- do a midyear conference, and then we'll kind of check in on if we -- how we sort of sense the trajectory. Philosophically, we don't believe that there's a type of customer that can't use Enterprise Advanced. So to reverse that, Enterprise Advanced is relevant for any customer. It's going to become the core of our sales motion and that will lead with Enterprise Advanced use cases. That does not mean that every customer can instantly elect to raise their hand for it. It is a premium price SKU in the ladder. You do have to be at a point where you want to use AI against more of your data from a workflow standpoint, but it's relevant for everybody. It will be relevant for a 10-person law firm, a 5-person boutique investment bank, a 50-person life sciences company. So there's nobody that's not relevant for, but it is priced as a premium product and for customers that are kind of ready to engage on that journey. And so that means it will certainly be an adoption curve.

And what does that adoption journey look like just from a timing perspective? Is it proof of concepts and -- or is it kind of a quicker upgrade? Like take me through that.

Well, so the deals we did at the end of Q4, we did a few dozen deals in Q4 as we called out in the earnings call. And that was super exciting, kind of exceeded our internal expectations. Partly, there's a certain mojo when you have a new offering. The sales team is kind of quickly going out talking to those early adopter customers. And so at most, it was a 2-month sales cycle for anybody from a standing start because we had only announced the product at BoxWorks in mid-November. So we wouldn't sort of typically see it as a 2-month sales cycle. In kind of normal course and speed, that was in that early adopter energy. But we're drafting off that right now. So now we're already, let's say, 4 months into educating the customer base. And so you're going to just see -- everywhere you look in Box, we will be promoting the capabilities in Enterprise Advanced. We'll be promoting, from a product announcement standpoint, the innovation that's coming out. We'll be promoting the customer stories out of Enterprise Advanced. So that will just shape and help improve that adoption curve over time.

Yes. And I'd say just to add to that, for the more normal and typical customer, we would expect to see pretty -- much more normal enterprise sales cycles and so to more fully show up in kind of the bookings at the tail end of this year and then to really start to have an impact on our revenue beginning with next year.

Okay. And from a monetization perspective, you've talked about 20% to 40% uplift from Enterprise Plus to Enterprise Advanced. I guess the question is, is there -- are there ways for your broader customer base to use some of the function -- like the new features in Enterprise Advanced separately add-on? Are there other ways to kind of get more of your base using these products? Or is it all about Enterprise Advanced from a monetization perspective?

Well, so we do have some API-oriented use cases that we think will transcend plan type over time, and we're still exploring how we'll expose those. And so maybe at some point, you have a slight decoupling of if you're a developer and want to use our APIs in a headless fashion for some kind of workflow, we want to make that available, which is maybe separate from your end user seats in your organization. That's a little bit kind of TBD. But I think we've -- I don't want to like overly lean into this point, but I think we're kind of nailing the SKU ladder. And in the sense of -- the best way to think about Box, and this has been maybe a year or so journey to get here fully. But by the end of this year, certainly, this will be the case, you should think about Box as if we had started the company in 2025, which is to say that -- and this is maybe distinct from other enterprise software companies, which is to say that we don't treat AI as if there's like a dumb version of Box and then a smart version of Box. If you were to start Box in 2025, we were just like -- we were just refounding the company, we would be an AI-first platform to work with your unstructured data. That would be the company. Like we would pitch a different version of the company. That would be all we do. Then you go to our pricing page and you'd say, okay, well, it sounds like for your base edition, you can do this. And then for the really advanced version, I can deploy agents to go and do much more automated things. That is where we're converging. And I think that throughout this year, you'll see more and more functionality show up at different plan tiers. So everybody will get a taste of AI. And as you move up the plan tiers, you'll be able to do more automation, more agentic workflows, integrate with other kind of AI systems and so on.

Great. Sticking to AI, I want to ask about AI agents and the AI Studio. Like any leading use cases? And I also want to tie that question around agents to your vertical strategy. I think there's a link. If you could talk a little bit about verticalization as well.

Yes. So the initial kind of AI agents that we've deployed one that is our kind of first maybe big one that will be a killer app is the ability to extract structured data from documents. And that happens to be big simply because it's an incredibly cumbersome process for organizations and enterprises already spend money on it. So it doesn't like take a leap of faith to understand why you use it. There's a category called intelligent document processing. Sort of its adjacent categories are like RPA and OCR or customers go and use like BPO firms to pull out data from their content. So if you go to a large professional services firms, they might be like reading resumes and pulling up the data from the resume and putting that into a structured database. Or if you go to a law firm, they're doing that with contracts. If you go to a bank, they're doing that with invoices and bank statements. AI is getting increasingly better at being able to do that, obviously, in an automated fashion. And so that's kind of the -- maybe the first killer app agent will be the ability to extract structured data from documents. Right now, that's only available in our APIs, but we've alluded to because of the acquisition we did with Alphamoon, that will show up in the product in a bit more of a kind of user-oriented way over time. So that's a really big one. And then I'll maybe kind of point to some hypotheticals over time of what you'll see within this fiscal year is we're seeing a lot more agentic workflows. So the ability to go and tell an agent, hey, I want you to look at these 10 documents and again, provide a full report on this. And so we have a lightweight version right now you could do if you were on Enterprise Plus. So after our earnings call, we have, let's say, 10 different analyst reports. You just click the 10 and you say, write a report summarizing all of this. And it will -- and you pick whatever model you want to use, and it will provide whatever type of summary that you want in any kind of format. You could say build a table, show me where all the price targets changed, what are the -- what are the negative and positive feedback, whatever you want to report to be. And that would be the kind of thing that you'd have an internal analyst go and do and spend a couple of hours on, and now it's 10 seconds of AI going and doing that. That -- those are the kind of use cases that you'll just see more and more of in every part of the business. So as users and as enterprises get more familiar with the fact that, okay, I can just like go talk to my data, you'll see this become just a way that people augment the way that they do their work with their information.

Perfect. I want to talk a little bit more about Box Apps, no-code and Crooze acquisition. I guess to start, what has that asset done to really open up a new set of TAM and use cases for you.

Yes. So I think this is -- it's kind of a sort of a sister to the agentic workflow side of things. So if you imagine the agents will kind of automate a lot of the busy work about what we do with our data, you often will still need an interface in a way to understand what work is happening. I'm a big believer that chatbots and you'll talk to things in the future that will be very pervasive. But there'll be -- there's lots of use cases where you still want just a dashboard that just like tells you what's going on. You don't want to every day have to log in and think, okay, show me how many customers are renewing this week. Like you don't want to type that query every single morning and then have it feed the answer. You just want to go to a page that says like here are all your renewals that are showing up. Box Apps lets you create those applications. So you have a contract dashboard. And the contract dashboard shows you everybody renewing, everybody who has out-of-date clauses, things that somebody else still has to go and review for approval. Previously, within Box, you just like literally couldn't have built that, right? We were a place that had folders, you put files in those folders, you'd have to know to go to the folder to find what you're looking for. With Box Apps, you literally go to a page and it looks exactly like a dashboard of any other type of software that was sort of purpose-built for its workflow. So you would have a dashboard for in life sciences, your clinical trials, like who's in the clinical trial, what stage are they at? Have they been approved? And that would be a dashboard as opposed to looking at a file system for that information. You are a loan processing agent. You would go somewhere and you say, here are the loans that are in review. Here's where I have to go check next as opposed to, okay, I have to go to a file system and find that document. So Box Apps is -- it's going to be a pretty significant breakthrough in what we can now deliver in terms of really powering that full workflow for our customers. And then agents become this sort of counterpart in that app as just letting you go and automate more and more of the workflows as a part of that application.

Great. I want to transition to go-to-market. First question is, is the sales force ramped and ready to sell all these new -- Suites, Enterprise Advanced, all these new AI tools?

Yes. I mean we are doing all of the mechanical things that you do to train, incentivize, enable, show the case studies, have the right presentations. We just had our kickoff a few weeks ago, certainly the highest energy kickoff we've had as far as I can remember. And so we're doing the things that you do to ramp a sales force and train a sales force. Will everybody have perfect efficacy right out of the gate? No. Some people will sort of have their stumbles and we'll work through that. But we're sharing the best practices. We're doing all the trainings internally. The QBRs are happening within the sales teams. So I expect this ramp to be pretty effective. Maybe a couple of points that maybe would be some areas that people would consider to be risky otherwise. This is not something that is like a meaningful change to like the kind of rep that you need. We've always been sort of selling part SaaS and application and part platform and APIs. So the reps that we have today, I have always lived in this middle ground of like pseudo infrastructure, pseudo business process. That's the only way you can do a deal at Box is you have to get IT sort of support, sometimes even infrastructure support and then you have to know the line of business use cases. Well, it turns out selling no-code apps and it turns out selling agents is kind of the same thing. Like what technology am I plugging into, what line of business value am I offering, you have to be able to sort of finesse that set of conversations and that flow. So the sales reps, I think, are well positioned to get there. We are building a broader partner ecosystem. So this is an area certainly that we're investing in, not in some crazy way, but in a methodical surgical way where we know that we need to get into many organizations alongside critical system integrators, the Deloittes, Accentures, Slaloms of the world as well as boutique system integrators that really, really kind of have dominated in the enterprise content management space. So it's a bigger kind of partner build-out that we'll be doing this year. And then maybe the final point as it relates to go-to-market investments is we need to ensure the awareness of what we're doing with AI. So if you are the AI person at literally any company on the planet, we need you to be thinking about Box whenever you're thinking about using AI and unstructured data. Like Box just needs to like rise to the level that it's on your radar. And so that will be incremental ways of getting the marketing message out there as well.

Perfect. Question on competition. We've been discussing how dynamic your market is around all the technology changes. So how has generative AI impacted your position in the market?

I think that it's -- I think the most -- so I think very positively. And I think the very exciting thing is that we see an opportunity to take a category that if you had said 2 years ago, let's -- just to make enough distance in time kind of pre-ChatGPT, 2.5 years ago, if you said like what do you think of the enterprise content management market? This is like not the sexiest space to be like thinking about. We've been doing it because we get excited about it, but it's -- there's some legacy players. It's not really changing that much. Maybe there's some workflow automation. It's -- we love it because we signed up for it. But again, you go to a Fortune 500 CEO or a CTO, and it's like not the top priority that they're going to be paying attention to. And so you kind of grind it out in the mid part of the business and so on. Today, you sort of say, okay, like where does unstructured data rank relative to what you can do with AI and the impact it will have, and it's really, really high up there as kind of previously discussed. So what's happening is it's going to sort of shock the system of this category, which is, okay, actually, it turns out that the places where you put your unstructured data become extremely important because if I have that data in a legacy system that is not AI-ready, and I'm pretty [indiscernible] because all my contracts, all my financial records, all my resumes, all my project plans are sitting in a kind of non-intelligent, non-AI-ready system. So that's going to cause a lot of customers to say, "Oh, I need to do an upgrade cycle," or it will cause the companies that never even invested in these kind of platforms to say, "Hey, I actually wasn't treating this as a strategic asset type. Now I actually can. Is there a platform that lets us go and actually work with AI and this data." So this raises our competitive position because we've compressed kind of all the things you need into one stack. We do unlimited storage. We have a nearly infinitely scalable file system. We do all the security permissions around your data. We support all of the major compliance requirements that you would have when you need to manage your data. We have an AI abstraction layer. So there's very few companies, maybe 0 that have that full kind of technology stack compressed into one thing. You could go to maybe a hyperscaler, but guess what? The moment that you want to use -- you put your data, let's say, into something like Google Drive or whatnot, and then you want [ Anthropic ] to work on that. Now you're building custom software to go and connect what you're seeing as innovation in that space back to your data. So the ability to have a single stack that's intelligent by design, connects to all of the AI models is a sort of end of one type platform architecture that we believe improves our competitive positioning, and we're seeing it in certainly the deals and win rates right now.

Perfect. You've had a strong partnership and integrations with Microsoft. And along the way, we always hear about customers deploying Box alongside Office 365. Just wanted to check what you're seeing as far as Box AI. Are you seeing that used in conjunction with Microsoft 365 Copilot?

Yes, Yes. I mean 100%. So we have an integration with Copilot right now, and it's sort of actually more orthogonal to Box AI. So Box AI is when we're doing the intelligence on the data, and then we have an integration with Copilot where they're sort of talking to our APIs, sucking data in and then doing the intelligence as well. So it's -- nomenclature-wise, it's sort of just a core Box platform integration. Where we think this industry is going to go, we're extremely confident about this part. And it's more of a shared, I think, understanding by most people in the industry at this point is you will need agent-to-agent communication between systems. So Copilot, let's say, their default core agent will -- you'll be talking to Copilot and you say, "Hey, I want to be able to summarize this earnings call or get a sales material." It will talk to our agent and then go pull that information from -- our agent will pull the information from Box, pull it out, summarize it, send it over to another agent and then you'll feed it in. That probably is the direction of this industry broadly, which is really just AI to AI type communication and integrations, and we would expect to see even more usage of that model likely in the future. But the partnership with Microsoft is great. We plug into Copilot. From our perspective, we want there to be as many places as possible that people are working in that talk to data because that obviously increases the need for a horizontal neutral platform that connects to all those technologies.

Thanks, Aaron. Dylan, let's talk about the quarter a little bit and FY '26 guidance. There were definitely a lot of moving pieces, both in Q4 and in guidance. I was hoping you could help to unpack or sort of provide the messaging around both growth and margins.

Yes. So big picture, Q4, very strong quarter for us overall with kind of outperformance kind of top and bottom line, including on the billings side, which was fueled by a lot of the early momentum adoption of Enterprise Advanced, which caused customers to early renew so they get access to those capabilities as quickly as possible. For next year, this year now, FY '26, say, big picture, kind of expecting it's an investment year for us because of all of the excitement and opportunity that we have around Enterprise Advanced, AI and just in a critical moment in time with this technology and the landscape evolving. At the same time, we are still committed to expanding our operating margin for the year, albeit at a rate that has been lower than we've seen over the past few years for us. So think about it as slight operating margin expansion to invest in growth. And the numbers include kind of cutting through all the noise of FX, leap years, all the other stuff, you can think about as generating consistent revenue growth, an acceleration in billings growth to 7%, which is a little bit ahead of our revenue growth expectations of 6% and starting to see some momentum that we're seeing even in the guidance numbers that we provided, whether that's our net retention rate improving by 1 point year-on-year, revenue gradually accelerating throughout the course of the year, again, billings being a little bit ahead of revenue. So that's kind of the big picture is we're investing for the future growth. That's not going to show up hardly at all in revenue this year, but excited and firmly on the path to return to double-digit growth.

Perfect. There were also -- there was also some movement on the non-GAAP tax rate. I wanted to give you an opportunity to...

Yes, it's a result of now that we've been consistently profitable in the U.S. for a few years and forecasting the same. There was a valuation allowance that impacts those non-GAAP numbers, noncash. And what we had talked about and disclosed was there is a 27% rate, long-term rate that we're using that is reconciling GAAP to non-GAAP items. But then when you factor in preferreds and the cash taxes that we mentioned, that will be pretty modest this year in the $12 million to $15 million range. If you think about the full non-GAAP tax rate, that ends up being about 32% -- so definitely a lot different than previous years just because of that U.S. release due to our profitability, but something that you can expect to see going forward.

Okay. And so putting in slight acceleration in billings growth as a proxy for growth, some improvement in margins, like how does that combination of growth and margins fit into your existing framework around growth plus free cash flow margin?

Yes. So I'd say philosophically, think about the 2 kind of in conjunction and developing and just really delivering profitable growth and seeing expansion both top and bottom line in the coming years, that is certainly nothing about the philosophy of things has changed. I would say that the actual improvement this year is lower than what you'd expect to see from us over time in terms of that combined, okay, how much is revenue growth plus kind of operating margin improving, we would expect at least a couple of hundred basis points per year for the next few years. And this year, again, because you get that we're making the investments, but it takes time to show up in revenue. And then based on how well those are working, you would see that more concretely in the revenue growth, and that's why we're investing more or, "Hey, revenue is going to only improve incrementally, but you're getting a lot more margin." You can think about it as a balance, but this year being pretty unique in terms of the actual impact of the expansion of those combined metrics.

Excellent. We're out of time. Thank you, Aaron and Dylan. I really appreciate it.

Thank you.