Our CTO, Ben Kus. The leadership team will provide an overview of today's exciting news, as well as do a deeper dive on Box AI. Our Chief Revenue Officer, Mark Wayland, will also be joining us for our live Q&A session at the end of the presentation. Feel free to send questions to ir@box.com or enter them into the text chat box. Please note, this presentation may contain forward-looking statements that involve risks, uncertainties, and assumptions. Further information on the risk factors that could affect our forward-looking statements we make in this presentation can be found in the documents that we file with the Securities and Exchange Commission. With that, I'll turn the session over to Aaron Levie, Box co-founder and CEO.
Thank you, Cynthia. I'm incredibly excited to be here today after we've announced some major updates at BoxWorks to all of our customers. Diego and Ben, in just a few minutes, will be walking through our AI strategy, as well as some of the other big announcements from the day. As we discussed in our Financial Analyst Day earlier this year, we have a long-term strategy for driving consistent, profitable growth at Box. We're attacking a massive $74 billion+ market with significant tailwinds. We're building the leading Content Cloud that powers critical workflows across our customers' organizations. We're going wider and deeper with our over 100,000 customers by increasing our Suites adoption, and we're driving sustained growth and higher operating margins to achieve Rule of 45+ target in our long-term model. We're excited to continue to be on track to deliver against these goals.
Today, we specifically want to deep dive into our focus on driving and building the leading Content Cloud for our customers. So again, we'll be sharing some of the recap updates of what we've announced at BoxWorks today, and diving in a little bit more on what we're doing with Box AI. At Box, our mission is to power how the world works together. We're incredibly proud to be able to deliver against this mission for over 115,000 companies globally. Everything from small and rapidly growing startups to some of the world's largest organizations, including 68% of the Fortune 500. The reason that all of these customers have chosen Box is that content is our most important data. Content is the lifeblood of an enterprise. It's the sales materials that close a big deal.
It's the marketing asset that turns into a major new ad campaign. It's the R&D data that lets us deliver new products. It's the legal documents that close a big deal. It's the HR documents that onboard employees or train them, and it's the operational information that allows us to drive efficiency in our organizations. The challenge that most enterprises are dealing with, though, is that legacy approaches to managing content today are fundamentally broken. We hear this all the time from our customers. If you look at their IT environments and you look at where data is going, it's going across a wide array of applications and infrastructure, document management systems, collaboration tools, legacy enterprise content management systems, and then point solutions like e-signature vendors or FTP sites. Data is just fragmented in so many different applications, and this is a massive challenge.
Not only does it lower productivity in the organization and cost way too much money for IT organizations, it also means that there's massive risk in the organization. What we found is that when companies have their data siloed, 96% of companies say they don't know what's inside of those silos, so they don't know what actual intellectual property and valuable information they have inside of that data. Then, when you combine this with the fact that most organizations are dealing with cybersecurity risks and the cost of a cybersecurity breach is $4.5 million on average, this is a massive crisis. We have data that is in fragmented systems, that's highly important information, and the cost of losing that data is incredibly expensive to an organization. So there has to be a better way to solve this problem.
We have to be able to manage our information effectively, make our employees more secure, and be able to protect our most valuable information. That is why we've been building the Content Cloud. It's the intelligent platform for managing the entire content life cycle and content journey in the enterprise. Everything from data ingestion and data protection to how we classify content, collaborate on it, automate workflows, get signatures signed natively, the ability to publish content or analyze and generate insights from our information, or retain and manage documents at the end of a life cycle.
That is what the Box Content Cloud delivers, and we extend the power of this platform into every application that our customers are using, from Microsoft 365 to Google Workspace to Slack, Salesforce, ServiceNow, IBM, WebEx, and every other application in the enterprise, as well as the applications that our customers are developing themselves. So that is the power of the Content Cloud. But we know that we are in the midst of one of the most important shifts in technology history. AI is driving a generational shift in how people work, and we're seeing this already every single day. We're seeing it in forms of individual productivity, with code writing or generating emails and documents, or being able to create presentations, the ability to review documents or get expert analysis across our business. That's the power of AI when it's brought to the individual worker.
The real impact is when we can bring AI to our entire organization, when you add up all of the work that we're doing with AI across the entire enterprise. This means we're gonna be able to build products faster, we can support customers instantly, we can personalize marketing, we can accelerate our supply chains, and reduce business risks. AI is having a massive impact in how we work. Nowhere is AI more powerful than when connected to our enterprise content. When you take the power of super intelligence that we now have with AI, and we connect it up securely and compliantly with our most important business information. This is what customers are asking for and what we see so much potential in. You know, when you think about it, actually, only about 10% of our enterprise data is structured data.
This is the data that's inside of our databases or our CRM systems or our HR systems, and we've always been able to understand what's inside of those technologies. We can ask questions about that data. We can run analytics on that information. We can pull out insights. But actually, 90% of our business information is unstructured data, of which the vast majority is our enterprise content. This is our marketing assets and our contracts, our business reviews, our product plans. All of that information has an incredible wealth of knowledge and business insights inside of it. And until AI, we haven't really been able to understand what is inside of all of this data and how do we use it to be even more productive in our enterprise. So AI will revolutionize how we work with our enterprise content.
It means that in sales, we can instantly get the right answer to a sales question to close that deal we're working on. In marketing, we can now tailor all of our assets to customers automatically. In R&D and product development, we can run through troves of customer feedback to make sure that we're delivering just the right product to the market. In legal, we can pull out risky information and clauses to make sure that we're closing the deal on the right terms. Across HR and our employees, we can enable employees to ask any question and ensure that they get an answer back instantaneously about the work that they're doing, or the enterprise that they work for. In operations, we can extract critical information from our data to automate workflows and business processes in a way that was never before possible with unstructured information.
AI will revolutionize how we work with our content, and it's going to transform everything that we do with our data. We are building AI into the core fabric of our platform so we can allow you to answer any type of question about your information, pull out new insights, automate workflows, discover answers to important questions across our entire corpus of information, generate new content, and have advanced ways of protecting data in the enterprise. Box AI sits at the center of the Content Cloud, improving and empowering every aspect of our platform. To share a bit more about what we're building and to go through some of the other major updates from BoxWorks, I'd like to introduce you to Diego, our Head of Product.
Thank you, Aaron, and hello, everyone. I'm Diego Dugatkin, Chief Product Officer at Box. Since we last spoke to investors at our Financial Analyst Day in March, we have continued to execute on our product roadmap. I'm very excited to share an overview of the innovation we are delivering to our customers, including fantastic announcements at BoxWorks. Let's dive in. As Aaron mentioned, Box is the Content Cloud, a single secure platform for managing the complete content life cycle. We believe content is a strategic asset for our customers' organizations, and there is tremendous value in enabling our customers to unlock the value of all their content securely in a single cloud-native platform. And with our platform, we have driven amazing customer outcomes for over 100,000 customers, from small and mid-sized organizations to the Fortune 500.
For example, Broadcom has been able to migrate 18 million files from their product lifecycle management tool to Box and cut infrastructure costs by $10 million through eliminating on-premise file servers. JLL, focused on commercial real estate, empowers over 80,000 employees worldwide with a custom real estate app built on Box. With Box, JLL has been able to drive 9x faster sales asset creation, reducing times from 3.5 hours to just 23 minutes. And Coalfire, a cybersecurity organization, executed the file migration of multiple terabytes of content from 6,000 SharePoint sites. Creating and publishing content to keep teams on the same page is important in a world of hybrid work and global collaboration, and you'll see in a few minutes, we're going to transform publishing with Box. Now, let's look at some of the other areas of innovation, starting with Box Canvas.
Canvas offers rich, interactive, visual collaboration and whiteboarding directly in the Content Cloud. This is an area of collaboration that hybrid and remote work has made incredibly important. Customers are really loving Canvas. We keep adding more and more capabilities every quarter, everything from quick-start templates to a new, shape and icon library, where feedback has been great, and we're thrilled with the response only three months into our general availability release. We have also enhanced Box Notes, our real-time collaborative editing tool, with additional user-friendly content creation, formatting, and organization capabilities. Now, it's even easier for teams to draft documentation like SOP manuals, create project plans, or do code reviews. We also enriched Box Sign by adding advanced template capabilities and custom branding and enhancing its extensibility via Sign APIs. And just this morning, we announced Box Doc Gen.
Doc Gen automatically generates custom documents on the fly from multiple data sources. Feed it a template, map data sources, and output professional documents in the format you need right in Box, ready for the next step in your workflow. We have an amazing roadmap building on Doc Gen to push further into workflow and automation. Don't forget, security is typically the foundation of why customers choose Box. We continue to double down on our flagship security capability, Box Shield, which offers deep learning-based malware detection and advanced data leak protection to protect content from ransomware and other attacks. This morning, we were also excited to announce a new integration partnership with CrowdStrike. With CrowdStrike joining the Box Shield ecosystem, there is no more advanced content security offering in the world.
No other platform helps you unlock the value of your content across more than 1,500 different apps, including Microsoft, Salesforce, and Zoom. Box is the only platform for managing all your content with all your people across all your apps. All of these innovations I just shared will help us keep expanding our vision. But as Aaron mentioned, we are entering the age of AI, and nowhere is AI more powerful and more transformative than with content. So let's turn into the main event and dive into the news we shared this morning around Box AI. You might remember that in May, we announced Box AI. No one knows content as well as we do at Box. There are billions upon billions of files in Box, and they're flowing inside and outside of our customers' organizations, and it's all happening in a secure and compliant way.
For that reason, Box is the AI platform for content and for our customers. Importantly, Box AI will be built into Box at an architectural level. Box AI will pervade all of Box and will bring highly relevant new experiences and services to our customers without sacrificing our enterprise-grade security and privacy. Initially, in May, we announced two Box AI capabilities in Box, both leveraging an integration with OpenAI. But because of our platform approach, we have already announced our integration to work and the intent to integrate with other providers of LLMs. Ben Kus, our CTO, will share more about our architecture and technology strategy just in a few minutes. But first, actually, let's go and focus on what our users can see and play with. The first capability we announced in May was the ability for Box AI to help you derive business insights instantly.
Users can ask Box AI about a specific file in Box and get answers right away, summarizing a document, for example, or uncovering the key findings inside a lengthy research study. This is an incredible experience and shows the content and the importance of why content plus AI is so special and why LLMs are such a major leap. As Aaron pointed out, content is our customers' most important asset because of what's inside their files. Customer data, financial information, key facts, and confidential content all are essential for our customers. LLMs allow us to interrogate and manipulate the data at scale like never before, and the possibilities are endless.
Of course, just as compelling as looking inside our files to find just what we need is the power to create new content, and with Box AI in Notes, our users will be able to create content in seconds with a simple prompt. Actions like draft a press release, outline action items from meeting notes, or generate personalized emails all are possible now with Box AI in Notes. So let's go back for a second to the Content Cloud. We're building artificial intelligence into the core of the Content Cloud so we can enhance and augment all of the content processes in your business with the power of AI. You will see Box AI continue to inform our collaboration products, but we'll also bring new capabilities to our security products like Box Shield and beyond.
Box AI will be able to orchestrate workflows, generate documents from metadata, and send them out for signature, and also retain them when they're already sent back, automating processes and helping people focus on driving new insights instead of managing rote tasks. We have a truly grand vision for Box AI, but today, BoxWorks and at BoxWorks, we announced the next two steps along our AI journey. The first one is a revolutionary new approach to managing content as a source of truth by collecting, curating, organizing, and distributing it all through Box Hubs. And the second is a set of Box AI APIs and developer tools that our customers can leverage to extend our AI capabilities to solve important use cases in their enterprises. So let me tell you both of these announcements, starting now first with Box Hubs. At Box, collaboration is in our DNA.
That means you can preview and share documents in Box, brainstorm or whiteboard in Box Canvas, create project plans in Box, and using Box Notes, and analyze engagements through Box Content Insights and more. Now, with the future of work becoming increasingly AI-powered, Box AI will be infused across all our collaboration tools. And today's knowledge workers need new tools to help them collaborate. So think about it. The amount data is produced by enterprises continues to grow exponentially, and this growth is showing no signs of slowing down. So as a result, employees are struggling. 62% of them say that they are spending too much time, almost two hours per day, searching for information while at work. This is a massive loss of productivity with very real business impact.
Put simply, keeping teams and organizations on the same page and making sure that everybody is working off the same accurate and timely information has never been harder until today. Box Hubs will revolutionize publishing in the enterprise, making it easier than ever to surface great content quickly and securely. Hubs reimagines the way content is created, organized, and published by breaking it free from older hierarchies and basically doing it securely into easy-to-use intelligent portals. Hubs will power a huge array of high-value use cases in the enterprise. It can be a single place to find sales materials, a brand library, or function as a team site for big product launches or events. And we've even added Box AI to power multi-document search, summarization, and content generation. Let's check it out through this quick video.
You're working with all kinds of amazing content, content that's much more valuable to your company when it's published for teams who need it. Imagine what your people could accomplish with fast access to sales files, training videos, brand images, and tons more. Introducing Box Hubs. Get ready for a new level of simplicity with intelligent portals for curating and publishing content easily and on your terms. Give teams company-wide the context they need to quickly find and make sense of their most valuable asset, AKA your content. And with the power of Box AI, get fast answers and insights that help everyone drive key results. Whether you're empowering event teams with a project portal, equipping new hires with an employee resource site, or building one easy place for all your brand assets, you drive alignment and productivity across your entire business by easily curating all your content with Hubs.
By the way, no coding required! Say you're gearing up to unveil a new enablement portal for your sales organization. You'll start by customizing a hub with your company's branding. Next, round up all your assets into an easy-to-understand content playlist. Then instantly publish your hub, securely sharing it with anyone in or outside your business. Even have Box AI deliver insights across multiple documents. Meanwhile, see how your content's performing across teams with easy-to-understand analytics, and do it all while keeping your content secure in the Content Cloud. Get ready to let Box Hubs be your guide and watch your content shine.
Let's dive into the combination of Box Hubs and Box AI a little more. Box AI-powered search takes that huge leap further by adding speed, scale, and relevance. Hubs has been deeply integrated with Box AI, and it can be conveniently accessed from the Box Hubs search field. Looking for a particular document? Our powerful engine will route your query to search. But if you want to dig and understand complex concepts around, you know, around and across the entire content of a hub, you can also ask Box AI any question and receive AI-generated answers such as summaries, information analysis, and more right away. And here is the real magic. With Box AI, you're not only looking into multiple documents with a hub all at once, you're also looking into the right set of documents all at once.
Because a hub consists of a curated collection of files and content that you can trust that Box AI is querying as a verified source of truth. This is a critical distinction in the enterprise, where getting the right answer to a question like, "What is our sales projection for next year?" or, "What is our contractual liability in this type of agreement?" can make a serious and tangible impact. Box Hubs really illustrates the concept Aaron shared earlier: nowhere is AI more impactful than with enterprise content. Our portfolio of collaboration tools is truly best-in-class and an incredible value to our customers.
As the world of work changes, our platform makes it easier for people to be productive and leverage an incredible set of innovations of AI, while also lessening the burden on IT and IT teams by enabling them to retire or replace point solutions and combat app sprawl. Box Hubs will be available in beta next year. Box's value as a source of truth for content in the enterprise has never been more valuable. This year, we've continued to add more value through our ever-growing set of integrations. Just for example, we connected Box folders to Slack channels and Salesforce objects, then linked them all together so teams can seamlessly work between both. We added Box Notes directly in Teams and enabled real-time desktop co-authoring across Microsoft 365 apps like PowerPoint and Word, with all edits automatically saved back to Box.
We also updated our Google Workspace add-on, which lets me attach my Box files directly to Google Calendar meetings. The spirit will continue with Box AI. Next year, we're bringing Box AI to Slack, so you will be able to query Box AI to ask questions of files in Box and receive your answers directly in Slack. Finally, we're bringing Box for Microsoft 365 Copilot to make it easy for your joint customers to use AI to power their Box content through interactions with Microsoft 365 apps, starting with Teams. Work will never be the same. As you will see today, there is so much more to come. These initial Box AI capabilities will be available in beta this November, while Box Hubs with AI will be coming out next year.
The second area of AI announcements today is centered on new additions to our Box Platform APIs and developer tools. As a reminder, Box is built as an open platform with powerful APIs, so developers can have tools to enable our customers to extend Box to their own custom applications and develop solutions to meet business-specific use cases and do it all securely. With Box AI, I'm happy to share that we will continue investing in our platform by adding new AI APIs and developer tools for those purposes. Specifically, we will be releasing a Box AI API to support text gen, Q&A within single and multiple documents, and also summarization capabilities. Through our APIs, developers will be able to integrate all the same cool end-user functionality we're bringing to Box in their own custom applications built on Box.
There is a wealth of use cases for customers and partners that can benefit from this. They will be able to build custom apps and workflows to analyze, summarize contracts, for example, or to help an analyst compile and extract data points from multiple documents to drive a research workflow. Now, all of the developer tools that we will be creating for Box AI are helpful, and I think of one as, one of the most notable ones is the Box AI Metadata Extraction API. This is really exciting. The Box Metadata Extraction API will allow developers to easily apply metadata to files at scale, so they can basically choose to auto-populate the metadata suggestions for specific fields.
Metadata extraction can be helpful for difficult-to-understand documents, such as legal contracts, and help extract fields for something a user might not be able to do because they're not expert in that area. It is extremely useful for files where there is a lot of data entry involved, like invoices, and it will save a lot of time for analysts who could be focusing on other value-added work instead. Now, we're just scratching the surface with how this combination of AI and our platform tools will transform business processes and workflows. You'll want to stay tuned into this area for next year. All of today's announcements further cement our position as the world's leading content cloud. This is only a fraction of the number of innovations we've shipped this year, and I could not be more excited for our roadmap.
The advent of AI and large language models is a powerful catalyst for content and Box, and how we are delivering AI is essential to understand. Our customers trust Box because no matter, you know, what new products we bring them, we always build them on a foundation of enterprise-grade security and compliance. That is especially important in AI, where there are so many important and new growing considerations around how data is shared, managed, and governed. Now, I'd like to have Ben Kus, our CTO, to tell you about how Box AI platform works. Thank you very much.
Hi, I'm Ben Kus, Chief Technology Officer at Box. We've seen how Box AI powers productivity, including text generation capabilities, answering questions on single and multiple files, plus how Box AI can be leveraged for third-party or custom applications. But with the rapid evolution of AI, this is just the beginning. Since we know that 90% of data in an organization is unstructured, it's very clear that most companies will be looking at AI to extract valuable insights from their content. And with the power of generative AI, there's a lot of technology options that would let you do this with either custom-built or off-the-shelf tools. However, with any application that accesses content, you have to be keenly aware of the security controls, and with AI, you have additional concerns, like: How do permissions work? What are the controls on model training?
What are the protections against leaking sensitive data? These questions are critical to avoid a disastrous situation where AI itself causes data leakage. This is why we built these principles into the Box AI architecture from the very beginning. Box AI uses existing Box controls to ensure that you can use AI with enterprise-grade security and strict granular access controls while isolating the models so they cannot leak the data. To do this, we have created the Box AI platform. The way it works is that when a query is made on a document, including all the queries that you saw in previous examples, the AI platform first checks permissions to ensure the user has access to this content. Then, the most important data related to your query is extracted using multiple techniques, including vector embeddings.
After that, the AI platform will augment the data to reduce hallucinations, including any custom instructions, and will ensure the AI gives the most accurate answer to the query. From there, the data is sent to the AI model, and then the response is returned to the user. Best of all, with the AI platform, you'll be able to standardize controls and technology for applying AI on your content, ensuring that you can use AI safely and securely. And this not only works on text-based documents. We will continue to evolve the power of the AI platform to support different file types, including multimodal content, audio-video files, images, and semi-structured data like Excel files and Google Sheets. Now, one of the most critical aspects of the Box AI platform is that it allows model flexibility.
We will select the models that are best fit for different use cases, but we will provide the capability to change the model to bring in a custom-trained model or a fine-tuned model that supports your specific needs. We know that different organizations will have different needs for customization and administration. So with the Box AI platform, you will have the complete visibility and control and a bird's-eye view over how your enterprise is using AI. With AI access controls, you'll be able to monitor and report on your organization's AI usage for compliance purposes. In the future, you'll be able to select from a curated set of off-the-shelf models and see how the answers to your queries might differ when making those selections. You'll also be able to designate custom instructions.
For instance, instructing your model to answer in different certain styles or never answer specific types of questions, such as legal or medical questions. Finally, from this unified console, you'll be able to build, test, and deploy your own intelligent apps to embed the power of Box AI in existing experiences. And of course, our AI platform will implement and respect our AI principles, which demonstrate our commitment to control, privacy, and transparency in this rapidly changing AI landscape. As Aaron mentioned, nowhere is AI more powerful than with your content, and Box AI is the AI content platform for future-proofing your AI content strategy. And with that, let's bring it back to Aaron.
Thanks, Ben and Diego. Those were amazing updates around Box AI. Now, I'd like to talk a little bit about how we're bringing Box AI to our customers. Generative AI is a nascent but high-growth vector in the industry. Most CIOs are exploring and expect to invest in this technology. We've seen that there's a massive CAGR that's occurring in AI investments. We know the vast majority of CIOs expect to increase or maintain their investment in AI. Yet, we know we're also early in this journey, with most customers still exploring new ways to advance this technology in their business.
So at Box, we've thought about what's the right pricing and packaging for this technology to make sure that we can bring it to the masses within our customer base. So we have a set of pricing, packaging principles around how we are enabling Box AI in the Box ecosystem. So the first is that we want to ensure that Box AI is a core part of the Box platform, not some separate add-on capability, that is sort of outside of the core platform. We wanna make selling Box AI seamless for our sales and go-to-market organization without requiring an entirely separate sales cycle that could be, you know, quite laborious or caught up with many compliance issues, that customers might have. We wanna provide further catalysts for customers to choose our higher-tier plans, like Enterprise Plus.
We know we have a massive opportunity to continue to drive an upgrade path for those customers. At the same time, we know that there's a variety of different use cases that customers are gonna have with AI, some of which might be, on, on an extreme end of usage patterns, or volume interaction that they would have with AI models. So we need advanced and additional monetization vectors, for some of these advanced use cases that customers will have. And we also wanna make sure we're containing and managing costs and targeting a gross margin neutral AI strategy. So we want to ensure that we're driving premium value for our customers, and they're paying for that value. So this is why we are introducing Box AI and some of the, the capabilities that you saw today into our Enterprise Plus plan for customers.
The way that this works is customers will have a certain allocation of end-user usage of Box AI for all of the productivity use cases you saw, things like being able to answer questions inside of Hubs, being able to answer questions inside of a document, or generate new content within Box Notes. We also know that there's gonna be a higher degree of consumption use cases, for instance, with our platform, or with some customers that might go over that end-user allocation that they were allotted. This is why we'll sell all, additional AI credits on a consumption basis to our customers, and that will be rolling out shortly. So the focus for now is: how do we ensure customers can leverage Box AI as a core platform component within Enterprise Plus, driving a major upgrade cycle for the Enterprise Plus plan?
We see a major opportunity to be able to go and drive the next set of customers in Enterprise Plus, leveraging AI as a core component of that. We've talked about the fact that 48% of our revenue in Q2 is from customers that have one of our Box Suites plans, where Enterprise Plus is the core plan that we're focused on. But we also know that 21% of our revenue is from customers who have our core Enterprise edition and at least one add-on product. So this is a very ripe market and audience right there that we can go and drive upsells in the near term from. And we also know that 31% of our revenue is from customers that don't yet have an add-on product today, but where AI is an incredibly attractive offering to help drive upgrading into Enterprise Plus.
Recently, we did a survey of our customers, and 96% of them expressed interest in Box AI in some form. So we see a massive opportunity to be able to drive that upgrade cycle, leveraging Box AI as a core driver of Enterprise Plus. At the same time, we know that there's gonna be more advanced capabilities that our customers require. For instance, maybe you'd want a premium model like GPT-4, or an equivalent model, or you wanna be able to have customization in the model parameters and the prompts that you give it. Or you wanna be able to do things like metadata extraction or more automatic activities with AI. This is why we'll be introducing advanced capabilities in a higher-tier plan targeted for next year.
In this all-new plan, we anticipate having very advanced AI features that, again, let you do model selection, maybe bring your own model if you have advanced capabilities that you need, the ability to customize prompts or have advanced administration and configuration, and much more. This is consistent with what we've been discussing with the financial community and our customers, which is we will continue to drive more advanced capabilities in the Box platform over time. As we have been driving that set of advanced capabilities into our platform, we've also realized higher price per seat as we've introduced these new plan types. Next year, we anticipate introducing a new plan that will not only include the even more advanced capabilities of Box AI, as well as additional functionality that helps our customers drive advanced content management, collaboration, workflow, and security needs.
Stay tuned for more information on this front. So that's the major updates around BoxWorks and Box AI. We're incredibly excited to be able to bring Box AI to all of our customers, and we're excited to share more updates with all of you soon. With that, we'll open it up for questions. Thank you.
Thank you, Aaron. As a reminder, please submit your questions in the text chat box on your screen, or feel free to email questions to ir@box.com. Let's assemble the team now, and execs, if you could please turn on your videos, and we'll begin momentarily. Our first question will be from Pinjalim at JP Morgan. "It seems like there's a limit to AI queries to 20 per person per month, with an additional 2,000 per company. Could that volume be another vector of pricing for Box?
Yeah, great question. And just to clarify, it's more framed as 20 included or allotted questions or interactions per user, as opposed to a limit. So the moment you go sort of over that allotment, that's when you dip in the pool, and then we will sell additional blocks of AI credits at the enterprise pooled level when you go beyond those 2,000. So we anticipate those generally being in sort of blocks of 10,000 AI credits at a time, kind of on a monthly basis that a customer would pre-buy for an annual subscription. And we'll be sharing some of the more specifics about that throughout, you know, likely Q4, as these you know as the system fully rolls out. But we've had overwhelmingly positive feedback on the pricing model from customers.
It really kind of perfectly balances customers that, you know, wanna be able to have an end-user productivity solution for AI that they can enable for their large employee base. But also recognizing that when there's power users, they wanna be able to, you know, go beyond a certain kind of average-user limit, and they recognize that they would, you know, likely need to pay for that just based on the cost, you know, of the infrastructure.
Thanks, Aaron. An additional question from Pinjalim: How are you pricing the Box AI APIs? Could Box AI APIs really supercharge your platform revenue?
Yeah. So, the Box API credits will actually be the exact same structure as the pooled enterprise credits. So, you can think about the, for instance, those 2000 credits that a customer will get in Enterprise Plus as being able to be used by either the end users that go over their allotment or via any of our kind of volume-based use cases, like metadata extraction or API volume. And then you'll be able to buy again in kind of... You know, we're anticipating units of 10,000 at a time, API credits or those pooled credits with a very kind of similar pricing model.
So we do believe that this will be a meaningful, you know, kind of platform opportunity for us, and it's one that we're actually seeing a lot of customer demand from. Ben, I don't know if you wanna maybe share a little bit about what we're hearing from customers on the platform front on AI.
No, I think you started to say it well. Customers are interested not only in using the AI in the product that you've seen, but also they have a lot of different projects in their custom portals and other areas to use. They wanna use AI in their content, and using Box AI through our API solves a lot of their problems. So then they're very interested in using Box for that, as opposed to having to build all of the different platform techniques themselves in using just a large language model off the shelf.
Yeah, I think that just to underscore that point, in Ben's slide, where he walked through the AI interaction with, with Box, you can think about that as actually, many technologies that a customer would on their own have to go and install, set up, manage, and maintain just to be able to do the first interaction with a single document. And it's what we've built in our infrastructure and, and effectively are offering that as a service via our end-user experiences, as well as APIs. So, offers a ton of value, by abstracting a lot of that complexity away from developers.
Thank you. Final question from Pinjalim: If users are able to query their Box data using Microsoft 365 Copilot and our Box integration add-on, could you help us understand what might be the incentive for a company that is not on the Enterprise Plus SKU to upgrade that SKU to use Box AI?
Yeah. So, it's a great question. So, you know, the way that we tend to think about it is Copilot is gonna be more of a horizontal productivity play. And its value proposition is to go, you know, through email, obviously content, CRM data, you know, within Dynamics and, and, you know, any other kind of application, as well as plug into third-party applications like Box, which we've announced an integration with them. Our job is to go really, really deep in content-centric workflows and use cases, so things like the metadata extraction, or being able to have any kind of advanced content use case, like, Box Hubs, where you're doing multi-document querying, as well as the ability to have a neutral model, where you can bring your own AI models to your data within Box.
So, we are extremely differentiated on the dimension of going deep in content and the content-centric use cases customers have, but we also know that customers will likely want to interact with their content inside of a Copilot experience as well. So, you know, we see this as much more complementary, as opposed to any kind of competitive dynamic. And I think our model of including Box AI in Enterprise Plus, and then certainly beyond, with future plans, is actually extremely attractive to customers, and particularly those that might be considering a Copilot or Box AI strategy right now. Simply because, from an economic standpoint, the Box AI approach is gonna be much more just economical and an efficient way to get it into your business.
So, I would, you know, suspect that we're gonna have a lot more of a seamless experience in getting Box AI rolled out, due to the underlying kind of business model that we've gone with as well.
Thank you, Aaron. I have more questions from George Iwanyc with Oppenheimer: Could you expand on how Box AI and partner AI tools work together? For example, the joint solution of Box AI with Salesforce Einstein GPT, and Box with Microsoft 365 Copilot. How are the large language models and models deployed? Where's the data stored, et cetera?
Yeah, so Ben, I don't know if you wanna maybe go through some of the architecture and how we're working with some of the AI partners.
Yeah, so, at a high level, when we're operating on using AI on the data, the data is in Box. And so we will take care of making sure that we securely access the data, and then apply the AI. Now, the AI models can come from multiple places, including some of our partners, and Azure, OpenAI, or Salesforce Einstein GPT, and these are options that you can use to then apply the AI on your data. Additionally, when we have integrations with our customers, like we do with our Salesforce integrations or our just all of the integrations, like with Office 365 and others, you can be able to use the AI.
The AI will work on your data, so if you were editing, let's say, in Microsoft 365, and then using Copilot to maybe make, let's say, your Word docs, the data would then be stored in Box, and you can then perform AI on the data there.
Maybe something to add there is that you can think of Box AI as a pre-processing brain that allows us to define which is the best LLM to use of a collection that we have. So while we partner with multiple implementations, also connecting with the prior question, Box AI empowers our customers to get the best answer from different specialized options. And the having a joint solution that works with Einstein GPT or with 365 Copilot, it's not just the limit there. I mean, we can work with many others and actually select on the fly where the best answer is going to come from, and also optimize for cost, if that's also another variable.
But particularly, optimize for effectiveness, reducing hallucinations and building out architecture that is guaranteed to pick, you know, what is most appropriate based on the question that we're getting or the task at hand. So the intent is to continue to build a very neutral approach to AI and not forcing our customers to pick only one of the answers, like Aaron said, that may be too horizontal or not specialized enough. So, variety, in this case, benefits the customer.
Thank you. We've had a similar question from George at Citi, as well as another investor, and that's really talking about the implementation or onboarding of Box AI. Exactly how much customization is needed? Who does it, and how long does it take, roughly?
So in general, we don't expect much of an onboarding requirement because we're embedding Box AI across the portfolio in a way that the product itself has a no-code requirement and is very straightforward. Of course, there is a set of APIs that would require, you know, development if you choose to develop using Box AI APIs. But if you're using the product as it comes out of the box, the implementation of Box Notes with AI, it has a simple icon that you can simply deploy without any particular requirements or even additional training. It's ultra straightforward.
Yeah, just to add a little bit to that. It's part of the very important to the Box AI platform is that it's be able to immediately be used with no customization out of the box, like Diego says. So you just click, and it starts to work on it. But the power that we discussed in the presentation is to show that if you want advanced customizations, you can do those, and then it will then apply to all across Box, be it in the product or via the API.
Great. We have a question from Brian Peterson at Raymond James. How early do you feel you are in terms of use cases for AI in Box? Are there some early use cases in particular that you think are most likely to drive people to the higher-priced plans?
Yeah. So, to me, I'll share some of my conversations. And Mark, I know you've talked to a bunch of customers on this, and Ben and Diego as well, if you want to add product color. I think the use cases that we're hearing are, you know, very vast and broad. Mostly because every domain has a different set of kind of critical value that they generate from documents or content. So if you're in finance, it could be things like equity research. It could be trading data. If you're in any kind of real estate, it's likely, you know, lease agreements and property data that you have. If you're in legal, it's contracts and filings.
So, if you think about, you know, what's inside of all of our unstructured data, it really depends on the line of business or the industry. The very powerful thing about Box AI is that it works across, you know, nearly every vertical, nearly every line of business, so each customer kind of comes with their own set of use cases. We're seeing some great end-user productivity use cases where people say, "Okay, I wanna, you know, simplify an end-user workflow or process." That's usually things like document summarization, asking kind of questions of data.
Hubs is certainly gonna open up a variety of those use cases because you'll be inside of something, maybe like a sales portal, and you would have a sales rep ask, "Well, what's the latest pricing that we have for a new product?" And it's now gonna just generate the answer for you instead of forcing you to go to a document to go look it up. So that's a lot of end-user value and end-user productivity.
And then I think we're hearing some of these more mission-critical use cases as well, where a customer says, "Hey, I have 1 million loan agreements or, you know, 1 million invoices, and I want to be able to ask a particular question of all of that data in a very kind of structured way, and then maybe even automate a workflow against it." And those are the kind of use cases that will be very platform-oriented, a lot of metadata extraction, and where AI is increasingly a very powerful technology to help us with that. So, I think it's gonna be pretty, pretty much across the board, but really these, these sort of high-value, mission-critical, vertical use cases, that we're, we're pretty excited about.
So maybe, Mark, over to you on kind of driving that Enterprise Plus upgrade cycle, and some of the initial, maybe the reactions or conversation on that front.
Yeah, just to clarify, I think the question was moving above Enterprise Plus, right, to a future plan, or is it-
Uh,
Enterprise Plus?
I think with... Yeah, sorry, I think with—
Both questions, actually. Sorry.
Oh.
Maybe we could just address both. We've had the question-
I'll hit both. Okay.
Yeah, yeah.
Planning upgrades.
Yeah.
We'll just do the current-
Yeah, so we're a couple, we're a couple of years into the Enterprise Plus migration path and, and have migrated up many, many customers. One of the primary reasons why customers move up to Enterprise Plus is for our Shield offering that is part of Enterprise Plus, so the advanced security capabilities that have now been out for several years. That's probably the primary lever that gets customers to move. And now we will have another really compelling lever to get customers to move with Box AI.
So that, 21% of our revenue that comes from, that Aaron shared earlier, that comes from customers that have our core products, plus at least one of our add-ons, those are going to be a cohort of customers that will be, you know, very much at the ready to move up, along with Box AI being in Enterprise Plus. And then in the future, to, to get that set of customers that are on, Enterprise Plus to move up to that new future plan that, that we shared today will be next year. We're still actually defining everything that will be in that plan. But the, the opportunities are somewhat endless, as you saw in Diego's presentation, we've really stitched Box AI into every part of the Content Cloud.
These initial use cases that we're launching around Preview and Notes, and the really exciting one around Hubs with AI, there are an unlimited number of capabilities that you can imagine when you think about all of our capabilities around metadata, around workflow, around forms, and doc gen. Then injecting the power of AI into all those platform components really opens up a lot of really powerful use cases that will motivate our customers to move up to that new plan.
Great, thank you. Erik Suppiger at JMP: "What are the economics of using OpenAI services as part of your service, and how much do you pay OpenAI?
I can take that. So, we're not talking through the specific financials on this call, but we went through some of the in the presentation. So the technique that we use that we laid on the AI platform and what we said in there was we expect it to be gross margin neutral or negligible overall. So, we'll be able to contain the costs within the normal envelope that we work in.
Thank you, Ben. I just do want to go back to the question: what do you view as the primary triggers to drive upgrades to the higher tier above Enterprise Plus?
Yeah, great, and, and Mark touched on this, as he talked about the Enterprise Plus upgrade. And as we, you know, discussed in the, in the keynote, we will be introducing another plan, above Enterprise Plus next year. And as Mark kind of discussed, you know, AI will certainly be very infused in that. There's a lot of advanced, use cases that we have, that we're working on. But we think there's a huge opportunity for more advanced content management, workflow, security, oriented use cases. So you know, we kind of alluded to some of the capabilities, that we're working on, things like document generation, some of our more advanced metadata use cases.
You know, certainly as customers have more data, larger file sizes, there's a lot more that they need to be able to go and manage that content. So, we'll be introducing that plan again next year, and that will include a variety of these more advanced features that also will continue customers on their AI journey as well with Box.
Thank you, Aaron. We had a follow-up question from Pinjalim Bora at JP Morgan. Are you able to talk about what's under the hood for AI in terms of the vector database or vector search partners that you are using? And would Box venture to create its own specific LLMs or fine-tune LLMs at some point?
In terms of the underlying infrastructure technology, we haven't announced which database we're using. That's, but we've been evaluating all of them. As you know, we have this extensive infrastructure around searching documents, and we are looking to use the similar techniques for our AI embeddings search overall. In terms of the large language models, we believe that, as stated, we are very open to the idea of that AI models are evolving and that many things are, many of the models that are available are good at different use cases.
And so we'll continue to evaluate that, both from our partners, in addition to the open source models, in addition to considering fine-tuning models overall. But for right now, the power of the models that we've announced is we consider to be quite good at the use cases that we have announced.
Thank you, Ben. We have another question from George at Citi: Based on early feedback from customers, can you help frame the ARPU uplift potential you expect from consumption-based pricing model? And is there any offset to the addition of consumption-based pricing in lowering the current base pricing? So I do just want to remind you, we're not specifically answering financial-based questions, but I will let Aaron maybe opine here, just theoretically on the strategy behind it.
Yeah, so on the first piece, on the ARPU, because it's consumption based, it really will be just determined by the customer's kind of use case. So it's hard to kind of think about it as an ARPU at this stage. We'll certainly have better visibility into that as we know the variety of use cases. So, and as you know, Cynthia mentioned, we're not kind of calling out specific financial changes. So, but I think we're going to see just a variety of use cases that customers have.
Some customers will probably need in the hundreds of thousands or millions of AI credits, and some might just go over their allotment by a limited amount and need just, you know, kind of a, you know, kind of a very, very tactical upgrade. I think it's going to be a pretty broad set of use cases there. I'm not totally understanding the offset of additional consumption-based pricing in lowering the current base pricing. I don't know. Cynthia, have you-- do you know, or can you clarify?
George, we can take that. We'll follow up offline just to get more clarification there.
Because I just want to be very clear, we're not lowering any current base pricing, so I just wanted... I didn't understand what the question was, so.
Thank you. Josh with Morgan Stanley: If Core pricing was 100, what does Core Plus and Suites equal? How much uplift is there going to be from a Core Plus customer that already has had one add-on to Suites. So I will comment, at our Financial Analyst Day in March, we did give some of the economics of Suites customers versus Core Plus customers, as of the March date of, our FAD presentation. I think, again, once again, I think this is something, Aaron, maybe you can just talk the strategy behind, driving customers to Suites and then E Plus.
Yeah. So, yeah, and we can certainly share the last metrics that we've shared, you know, offline for what that kind of Core Plus cohort to Enterprise Plus looks like from a pricing dynamic standpoint. But yeah, the strategy is, I think, you know, well understood by the community. But we've been on this steady march of getting customers from our Core plan to have add-on products, and then with our Core, that kind of puts you into the Core Plus category. And then move customers into Suites, which Enterprise Plus is the kind of core focus on.
And so as we noted in the you know kind of keynote today, we see a massive cohort that is on Core Plus, so that means they have at least one add-on product, and we can move them up into into Enterprise Plus. That has you know usually a meaningful price you know uplift as a result of moving up to the Enterprise Plus tier. And then we think that AI will become another catalyst to have the Core-only population move up to a higher-tier plan. We've been on a meaningful number of conversations with customers that are on our Enterprise edition of Box, maybe not without... with an add-on product, but AI is the next thing that they need to get solved in their organization.
They have a mandate, you know, usually from the board or CEO on down, that they need an AI strategy, and one of the kind of most obvious areas where AI is gonna impact their business is on their data and their unstructured content, and so Box is the very relevant platform offering to go solve that problem for them. So we think that's gonna be a catalyst for Enterprise Plus upgrades. Mark, I don't know if you want to, you know, kind of comment on any anecdotal conversations or what you're seeing from the field on that front.
Yeah, it's, I mean, as Aaron said, so from a Core-only customer, it's generally a bit more of an uplift to get to E Plus. If they have our Core offerings with one or two SKUs, then it's an easier move for them to make, because they've already sort of partially paid for the Suite because they've bought some of the add-on products that are in it. So, you know, and like I said earlier, the big lever that generally gets our customers to move is Shield, and now we have another very meaningful lever to get them to move. So we're expecting to have another wave here following up from the announcement of Box AI, especially Box AI Plus Hub.
We're gonna have another wave of E Plus upgrades, which is gonna be very healthy for our business.
Yeah, Mark, from, you know, maybe just the past maybe month or two of customer conversations, anything coming up from an AI standpoint, that you know you think is gonna cause that upgrade motion?
Well, I mean, it's well understood that this is a technology wave unlike any that we've seen before. Two nights ago, I was hosting a panel at a CIO event. I'm in Europe right now, so there's a table of 20, you know, large enterprise CIOs from major European companies that you've all heard of, and it's 20 out of 20 that have AI projects and AI teams that they're building right now. And this is even different, you know, from my experience from the birth of the Internet itself, where, you know, as late as 1999, if you had 20 CIOs from large companies at a table, not all of them had Internet projects underway. Some of them thought, "You know, we're not an e-commerce company.
This doesn't apply to us for anything other than directions to our office." This is very, very different, where it's 100% of companies know that this is a mission-critical thing for them to pay attention to, and it is without regard to industry, public sector, private sector, financial services, life sciences, consumer, every single vertical, every industry all around the world. And they are definitely trying to figure out: What are we gonna build? What are we gonna buy, and who are our strategic partners going to be? And the feedback that we're getting is very interesting because these new generative AI models, these LLM models, are particularly well suited for text-based content. And here at Box, we are the stewards of arguably the largest single content store of text-based content on Earth in one multi-tenant cloud.
So it just unlocks sort of unlimited possibilities for us to create an avenue for enterprises to get access to the intelligence and the 90% of their data that's locked up in unstructured content. And that message really resonates, and it is very, very complementary to, I would say, the Copilot message from Microsoft of making, you know, office work in a much more intelligent way, which is a no-brainer thing for knowledge workers. I think that this is a very, very complementary approach, and that's the feedback we're getting from CIOs.
Awesome. Thanks, Mark. Maybe, maybe, Ben, I'm just gonna pick on you for one second because, you know, you, you've been leading a lot of the design partner program efforts, dozens, if not 50-100 customers at this point, that, that you've worked with on this front. And I'd love for maybe if you, if you can kind of approximate the journey that customers go through of like, what are they looking for? What do they think they need to do before the Box conversation? And then after the Box conversation, what tends to, to kind of happen?
Yeah. So, like I said, we've been on a tour talking to dozens and hundreds of our biggest customers to understand their AI strategy. And a lot of them start with the idea that they know they need to-- that AI is very powerful in new ways that they're considering. And, and, and on the top, maybe one or two, is the idea of their content. And because this is the power of the new AI technologies, it can actually understand this in ways that weren't really possible before.
As they go, sort of, think this through, they start to try to analyze the different use cases to helping their employees be more productive, to helping them be solving problems, automating things that before were maybe like a very manual process. So, almost all of them go through a process where they start by looking at it and saying: "Oh, maybe I can use build something myself." Still, they start to run into some of these challenges. Like we talked through, some of the permissions challenges is often the first one you run into. Like, AI itself can learn something if you're not careful, or it can actually just like become a data leakage source.
And so then, after they try to think about how to solve those kinds of problems, and they think through the compliance challenges, the security problems, and they then start to think, oh, they need an AI platform to be able to then solve these kinds of challenges overall. And of course, that's exactly what we're talking about, because we're looking to not only solve those in a standardized platform way that is both easy to use but then customizable, but then also, the important part of many of our conversations is what will happen next? What will happen as the models increase and the multi-model, multi-modal models become more powerful? What'll happen next when the AI gets better at certain things, or when it starts to have more advancements come?
And then this is very important for the customers that we're talking to, which is to say that they can't be... Every customer can't build their own set of capabilities on top of this. This is why they like to customers like Box into our platform, and we believe this pitch is very much resonating from the perspective of: We'll take care of all the details on how to use AI in your content, and then the customers will be able to use it ongoing to then solve the problems of their organization that AI can provide. That's possible now, that wasn't really possible a year ago.
Awesome. Cool. All right. Thanks, Ben and Mark, for those anecdotes, and thanks, Diego, for the other questions. Cynthia, I'll hand it over to you.
Thank you. Great session, everyone. Thank you again for joining us today. A copy of today's presentation will be posted on the IR website momentarily, and we look forward to speaking to you again on our next earnings call. Have a great day.