Table of Contents
After two decades building Box into the backbone of enterprise content management, Aaron Levie sees AI agents as the key to unlocking trillions of dollars in dormant data value. Here's his playbook for the next platform shift.
Key Takeaways
- Enterprise data has a lifecycle problem—it's hot for days or weeks, then becomes dormant despite containing massive value for future decisions
- AI agents will automate content-oriented workflows across every enterprise function, from legal contract review to financial analysis
- Three categories of AI startup opportunities: incumbents asleep at the wheel, innovator's dilemma business model disruption, and net new use cases
- Agent pricing will eventually converge to software margins due to competition, but the real opportunity is uncapping usage beyond human headcount limits
- Neutrality becomes a competitive advantage—being model-agnostic lets enterprises access breakthrough capabilities as they emerge
- Starting young and maintaining curiosity across diverse use cases enables 20-year entrepreneurial runs without burning out
- Tech's political shift rightward was predictable, but Democrats lost the narrative on core issues like housing and climate technology deployment
The Dormant Data Problem: Trillions in Untapped Value
Aaron Levie has spent nearly two decades watching the same pattern play out across 115,000 Box customers: companies create valuable content, collaborate on it intensely for a brief period, then let it disappear into digital archives where it never gets touched again.
"You think about what you mostly do with your files—you create one, you share it, you collaborate, and then it goes somewhere and you never see it again," Levie explains. "That data starts out hot for the first couple hours, couple days, week or two, and then it goes into a place where you just manage it."
The economic waste is staggering. All those financial documents, contracts, marketing assets, and employee records contain insights that could drive product discoveries, make sales reps more effective, or help new employees ramp up faster. But without AI, that information remains locked away.
"Yet it actually contains an incredible amount of wealth of value for an organization because it might have an insight that would lead to your next product discovery," Levie notes. "But you have not been able to tap into it."
This is Box's moment. After 20 years of building enterprise trust and managing sensitive data for 67% of the Fortune 500, they're positioned to unlock value that was always there but never accessible. "For us, AI is just this massive breakthrough which is we can finally actually open up the value of all of this data that organizations are sitting on."
The Agent Vision: Millions of Content-Oriented Workflows
Box's AI strategy centers on creating agents that excel at content-oriented workflows—the fundamental tasks that drive enterprise operations but have never been automated effectively.
"We imagine within the Box environment that you'll create agents that are really good at content-oriented workflows," Levie describes. "That's a legal assistant reviewing a contract for clauses that you don't want to agree to. That's a procurement assistant that is reviewing invoices or payment terms and then automating some part of that process."
The applications span every enterprise function:
- Financial analysis: An agent that analyzes 20 financial documents from earnings cycles, identifies semiconductor industry trends, and generates investment recommendations
- Legal operations: Automated contract review that flags problematic clauses and negotiation points
- Marketing automation: Agents that extract data from digital assets and automate campaign workflows
- Procurement: Invoice processing and payment term analysis with automated approval workflows
The key insight is that these aren't standalone point solutions—they're components of a broader platform that gets smarter as more enterprise data flows through it. "We'll have millions and millions of those kinds of agents that get created. They'll help you automate the work that you do with content in every field, every industry, in every segment of enterprise and public sector."
The Multi-Agent Future: When Platforms Talk to Each Other
Box won't be the only place where work happens, and Levie knows it. The real breakthrough comes when agents across different platforms can communicate and collaborate.
"Work happens in Salesforce and work happens in ServiceNow and work happens in Slack and work happens in Workday and 500 other technologies," he acknowledges. "We're going to eventually need agents between these platforms to talk to each other."
The vision is comprehensive cross-platform intelligence. You might ask ChatGPT to run a report on a customer, and it automatically pulls data from Box, Salesforce, ZoomInfo, and other systems to construct a complete picture. Or you might initiate an ITSM workflow in ServiceNow that automatically calls out to Box for relevant documentation.
"Our agents are going to eventually have to talk to each other," Levie explains. While the industry is "unbelievably early" on inter-agent protocols, the foundation is being laid for this connected ecosystem.
Where Startups Can Win: Three Categories of Opportunity
Despite incumbents like Box having structural advantages in data and integrations, Levie sees massive opportunities for startups across three categories.
Category 1: Asleep at the Wheel "Most large companies today exist because incumbents were asleep at the wheel," Levie notes, citing Netflix vs. Blockbuster and Amazon vs. Barnes & Noble as classic examples. "You could underwrite probably a trillion dollars of startup opportunity simply because of the asleep at the wheel dynamic."
Category 2: Innovator's Dilemma The business model disruption is particularly acute in AI because it shifts from seat-based pricing to consumption-based models. Customer support software exemplifies this: "I used to sell 500 seats of software to 500 customer support agents, and now we have a world where AI is going to start to automate more of that."
Incumbents face a painful choice: cannibalize their recurring revenue or lose to startups with better economic models. "Nobody in the management team wants to have less revenue the next year and have Wall Street hate them."
Category 3: Net New Use Cases The biggest opportunity may be in areas where there's no incumbent at all—work that humans never got around to doing because it was too expensive or time-consuming.
"We have a lot of things within Box that I'd like to do that we don't have people doing right now, and we might be willing to spend millions more dollars of just net new spend," Levie explains. "It's not just a cost-saving endeavor—we should be generating new work with AI."
Code generation perfectly illustrates this dynamic. "All of that spend on AI coding is sort of net new spend in the software stack. Every company and every individual just has now decided to swipe their credit card and now buy AI to augment how they work."
The Pricing Evolution: From Labor Economics to Software Margins
One of the most interesting dynamics in AI is pricing. Initially, agents can charge based on the human labor they replace—often a quarter of a person's salary, which is dramatically higher than traditional software pricing.
But Levie predicts this won't last. "If I had to bet on does AI remain comped at labor or does it remain comped at sort of infrastructure cost plus software and some margin, I'm betting on the latter just because competition almost sort of is going to cause that to happen."
The logic is simple: If a human task costs $100 an hour and AI can do it for $50, someone will offer it for $40, then $30, then $20, until it converges to normal software margins.
The real opportunity isn't charging labor rates—it's removing the cap on usage. "Right now when you're selling seats of software, you can only sell the number of people in the company. But now with AI, that 20-person company could have 10 AI lawyers and 10 AI SDRs and 10 AI marketers."
This creates massive TAM expansion. Companies will spend far more on software than ever before because they're no longer constrained by human headcount.
The Neutrality Advantage: Why Being Model-Agnostic Wins
One of Box's key strategic advantages is neutrality—they're not tied to any specific AI model or cloud provider. This matters enormously as the AI landscape evolves rapidly.
"When your content is in Box and one day there's a new model from Gemini that's really breakthrough, it just works. The next day there's a new OpenAI model, it just works. The next day there's a new Anthropic model, it just works," Levie explains.
This flexibility becomes crucial for enterprises that don't want to be locked into one vendor's AI roadmap. "I'm not stuck with 'Oh, this is only going to work with the model that they choose to deploy.'"
Box has made neutrality a core strategic decision. "We're a neutral platform, so we should probably take that to the max of all the software that we're going to work with, the models that we're going to support. We want to be just a neutral open platform, first mover, the first company to support any new technology for our customers."
Architecture Decisions That Compound Over Decades
Looking back on 20 years of building Box, Levie's biggest insights center on early architectural decisions that seemed minor at the time but compound dramatically.
One crucial choice was remaining steadfastly cloud-native when customers demanded on-premise deployments. "For obvious reasons, customers would say, 'Hey, could we run you on-prem?' And we were just very stubborn and steadfast and said no. We're a cloud company. It's all multi-tenant, it's all SaaS."
They lost customers in the short term but gained massive long-term advantages. "I'm extremely thankful for some very religious architecture decisions that we made early on because now every AI capability we add to the platform instantly works if a customer turns it on for the entirety of our customer base."
The lesson applies beyond Box: early technical decisions that seem costly often become massive competitive moats years later. "You don't have to be on version 19 of Box—Box is Box, everybody's on the same exact version, and when you have AI it just plugs into everything you're doing."
Staying Energized for 20 Years: The Platform Advantage
One of the most remarkable aspects of Levie's journey is maintaining enthusiasm for two decades as a public company CEO. Starting Box at age 20, he's now in his early 40s and more excited than ever.
"It only works because I enjoy it," he explains. "What we all did growing up and then eventually getting into tech and into startups is you like to build things, you like to create things, you like to solve problems, you get excited about new technologies."
The key is having a platform with range of motion. "We never got stuck in one vertical or one type of use case where you're just grinding out just incrementally improving that one use case. We've always had some range of motion."
Box's diversity of use cases keeps it interesting: "We help NASA go to space, major movie studios make a film, we help the research process of a new breakthrough drug. You can be a little bit ADD of all the things that you're doing."
The AI wave has reinvigorated everything: "I would be 20% less excited three years ago, but AI is just like holy shit. The demos that I see now on almost a daily basis from the team are just absolutely crazy shocking."
The Political Divide: Tech's Rightward Shift and Democratic Failures
Perhaps unexpectedly for a Box earnings call, the conversation turned to politics—specifically how tech has shifted rightward while Levie has remained on the left.
Levie sees the shift as predictable rather than surprising. "The left has moved left on a number of variables. Depending on who you are and what you're impacted by, those variables have become fundamental."
He understands why many in tech switched sides: issues like over-regulation, building restrictions, and cultural conflicts that directly impact technology companies. "Elon at some point was just like 'We can't go to space, I can't build things, we're regulating everything, there's a culture problem that I have a problem with,' and so obviously then you have to eventually switch parties."
But Levie believes Democrats lost the narrative on core issues rather than having inherently wrong positions. Living in California provides daily examples: "We live in California. It should be like the greatest place on Earth on every dimension. You have all the venture capital, you're sitting on this incredible asset, and then literally you can't make it affordable to live here. That's just insane."
The fundamental problem is execution, not ideology: "Democrats can't out-message that with their policy views because their policy views are in many cases just the wrong policy views. You actually just have to build and you have to create an environment where you can build things."
The Pro-Tech Opportunity in Government
Despite his political disagreements, Levie is optimistic about tech policy under the current administration. "There is a very pro-tech, very pro-innovation, very pro 'let's advance in a number of categories' set of individuals around the Trump administration in key cabinet positions."
His priority transcends party politics: "I care more than anything just about America continuing to be the best place to build companies, to build technology, to drive innovation." He sees alignment on AI policy specifically: "The AI messaging coming out of the government aligns more to my view of where we're at in AI right now—we just need more progress, we need more attempts at innovation."
The key insight is that tech's relationship with government doesn't have to be partisan. "This closeness that tech and the government have now couldn't have happened under a Democrat," but there's no inherent reason it couldn't have.
The Next 20 Years: Building on Compound Advantages
As Box enters its third decade, Levie is running the company as if he started it in 2025. "We're asking ourselves: are we doing everything as if we were starting from scratch in this new era of AI?"
The advantages they've built compound in the AI era:
- Trusted relationships with 67% of the Fortune 500
- 20 years of enterprise content flowing through their platform
- Model-agnostic architecture that works with any AI breakthrough
- Multi-tenant cloud infrastructure that instantly scales new capabilities
"We're in the right place at the right time in terms of having built out a platform that 115,000 customers trust," Levie reflects. "Now AI kind of plugs in right at the core of everything we're doing."
The next wave won't be about replacing Box or any other incumbent—it'll be about radically expanding what's possible when dormant data becomes intelligent, when agents can coordinate across platforms, and when the constraint on enterprise software shifts from human headcount to imagination.
For founders and investors watching this unfold, Levie's journey offers a crucial insight: the best time to build for the next platform shift is often decades before it arrives, when you can construct the foundational infrastructure that becomes invaluable once the shift occurs. Box spent 20 years becoming the trusted repository for enterprise content. Now AI makes that content infinitely more valuable.
The companies that win the agent era won't necessarily be the ones with the best models—they'll be the ones with the best data, the deepest customer relationships, and the architectural decisions that anticipated a future most people couldn't see coming.
