The Power of Data: Box CEO Aaron Levie joins OPTO Sessions

Aaron Levie founded Box [BOX], an enterprise cloud company, in 2005 while studying business at the University of Southern California. Today, the company boasts a market cap of nearly $5bn and provides content cloud services to approximately 67% of Fortune 500 companies, with Levie serving as its CEO.

Levie is currently supporting Box’s move from cloud computing into agentic artificial intelligence (AI) — autonomous AI systems capable of achieving tasks independent of human supervision. At its developer conference BoxWorks in mid-September, Box launched a new set of features allowing clients to build AI agents within the company’s products. 

But beyond building agents, Box’s success derives from its greatest resource as a cloud provider: data. In a world where training makes the difference between a successful use case and AI slop, data is key. “We are living in an era where your data is the most important context for AI,” Levie says.

In the latest episode of OPTO Sessions, Levie discusses how Box is rolling out AI tools, and how its preexisting systems are helping to deploy them at scale. 

Watch the full episode on YouTube, or listen on Spotify or Apple Podcasts. 

Unlocking AI’s Potential

AI is widely misunderstood, as industry leaders often point out. However, Levie sees that many of Box’s clients are eager to discover the capabilities agentic AI can provide them with. 

“Customers are more and more excited by that and learning about what that could mean for their business.”

Until recently, enterprise cloud companies have mainly focused on storing, organizing and protecting client data. For Levie, that represents a missed opportunity. “Enterprises have just a tremendous amount of corporate data in the form of contracts, marketing assets, financial documents and invoices. All of that data contains incredibly rich business information, but you’re never really pulling out the real insights from that data.”

For many of Box’s clients, the majority of their data is unstructured — “the messy, more human readable kind of data.” Technological constraints have largely kept this data untapped. “Computers in the past have never been able to really process 90% of the data that we work with,” Levie explains. “They’ve never been able to ask all of that data questions and get an answer back.”

By reading through massive amounts of data, AI agents, such as Box’s Box Extract product, can help enterprises identify what’s useful and what’s not. And with massive amounts of data in its systems, Box is uniquely positioned to deploy agentic AI at scale for its clients. 

Of course, a wealth of data is not enough — you need the systems in place to ensure clients’ data is safe, and that deploying AI agents will not introduce security risks.

“You need security, compliance, governance. You need to have a lot of scalability. That is the price of entry into being able to do AI on data.”

Box already has systems in place to ensure data is shared and accessed in a controlled manner, Levie notes. “One of the benefits of having your data in Box is that you set very clear access permissions and access controls for your data. If you create an agent, that agent can only access the data that I’ve shared with you or that you shared with me.”

Structural Advantage

Beyond accelerating workflow, AI has also supported the proliferation of opportunities for companies. “We are executing at the rate of a startup right now,” Levie explains. 

Box has largely taken a model-agnostic approach to implementing AI, with Box AI offering both Claude and ChatGPT models to allow clients to build agents. This is made possible by the model context protocol, or MCP, an open-source, open standard framework that allows AI systems to integrate with external tools or data sources. 

This helps solve the complex problem of interface between an external agentic system and Box’s data. “Without ever having seen our API before, our MCP server effectively helps that AI system understand what tools to use within the Box environment.”

Box’s stable financials are helping to support expansion of its AI tools, with Q2 2026 seeing a 9% rise in revenue year-over-year, and a 16% rise in remaining performance obligations. The company’s steady net retention rate has topped 100% for the last five quarters. For Levie, the reason is clear: “It’s the stickiness of the platform and the stickiness of these use cases.” Additionally, AI tools are driving demand for Box’s enterprise plans, with the percentage of revenue derived from its Suites offerings rising to 63% in the latest quarter. In short, demand for these tools is “causing customers to drive an upgrade cycle into our most advanced features and our most advanced capabilities.”

Ultimately, Levie sees Box as being uniquely positioned to build out these tools at scale, turning agentic AI into a reality for clients. “To get that AI to agents, you need a set of systems and tools that make that very easy to do and make it secure and make it well governed. And we’re one of the only platforms that can do that at scale with all of the critical business content that drives how agents operate.”