Box Unveils AI Automation Service to Tackle Millions of Invoices as CEO Warns Job Displacement Fears Are Overstated

Box is moving aggressively to position itself at the center of enterprise artificial intelligence workflows, announcing a new service designed to automate some of the most tedious and time-consuming back-office tasks that clog corporate operations.

In an interview at the Reuters Momentum AI summit in New York on Monday, Box Chief Executive Officer Aaron Levie told Reuters the company would launch Box Automate within 24 hours. The platform is engineered to let AI agents—software programs that perform tasks with minimal human direction—plug directly into business processes and handle massive volumes of unstructured data, from invoices to contracts.

“Box Automate lets a customer specify how AI agents can plug in to business processes and handle unstructured data like invoices,” Levie said. He added that the software could have an AI agent process 10 million invoices and pull crucial data from “every one of those invoices.”

The new offering builds on earlier AI programs Box developed to help enterprises harness sprawling and often disorganized data repositories. By zeroing in on high-friction, document-heavy workflows, Box is betting that companies will pay a premium for automation that can eliminate manual data entry and accelerate decision-making.

The company is hopeful that Box Automate will serve as a powerful upsell lever, steering customers toward its Enterprise Advanced plan. That tier allows organizations to build the custom AI agents that power the automation engine, potentially increasing Box’s average revenue per user in a competitive cloud content management market.

The ‘Last Mile’ Problem

Despite the sweeping automation capabilities, Levie struck a notably cautious tone on the broader narrative that AI will rapidly eliminate human jobs. In a separate post on X, formerly Twitter, he described what he called a “Gell-Mann amnesia” effect—a phenomenon where people recognize the deep complexity of their own jobs but wrongly assume AI can fully automate other professions.

He argued that workers who use AI in their daily roles see the limitations firsthand. “We run into issues about accessing data,” Levie wrote, noting that employees must carefully verify “the output of the AI” before it can be used effectively. He also pointed to the extra effort required to “incorporate that work into some broader business process.”

This “last mile” of human work, Levie suggested, involves context, judgment, and integration that machines still cannot replicate. Yet, he wrote, “We tend to dramatically underestimate the work that goes into making the AI work just as effectively in those jobs.”

A Broader Industry Debate

Levie’s comments land amid an intensifying debate among technology leaders over AI’s ultimate impact on the labor market. The discussion has produced starkly divergent forecasts.

Netflix co-founder Reed Hastings has argued that AI excels at structured tasks but remains limited in fields requiring emotional intelligence and cultural nuance, suggesting future education may pivot more heavily toward human-centered skills.

Khosla Ventures founder Vinod Khosla, by contrast, has predicted a far more disruptive shift. He has said AI and robotics could perform up to 80% of jobs by 2030, slashing costs while dramatically boosting productivity—a vision that implies a wholesale restructuring of the workforce.

Billionaire investor Mark Cuban has likened AI’s rise to the personal computer era, a moment when workers gained powerful new learning tools but also faced pressure to adapt quickly as companies streamlined headcounts.

Box’s product launch and Levie’s commentary sit at the intersection of these competing visions. Box Automate is designed to remove human labor from rote processes at an industrial scale—10 million invoices is not a niche use case—while Levie’s public remarks emphasize the irreplaceable human judgment that surrounds even the most advanced AI systems.

Strategic Implications

For Box, the automation push is a critical plank in its effort to differentiate itself from larger rivals like Microsoft and Google in the enterprise content and collaboration space. By layering domain-specific AI agents onto its existing content management platform, Box aims to become the system of record not just for storing files but for acting on the data trapped inside them.

The Enterprise Advanced plan, which includes the agent-building capabilities, represents a higher-value subscription tier. If Box Automate succeeds in demonstrating clear return on investment—by cutting invoice processing times, for instance—the company could see improved retention and expansion rates among its largest customers.

At the same time, Levie’s public caution on AI-driven job losses may serve a dual purpose. It reassures enterprise buyers that Box’s automation tools are designed to augment human workers rather than replace them outright, while also positioning the CEO as a nuanced voice in a debate often dominated by extremes.

The launch comes as enterprises are inundated with AI product pitches and are increasingly scrutinizing whether new tools deliver measurable productivity gains or simply add to the noise. Box’s bet is that a focused, high-volume automation use case—processing millions of documents—will cut through that skepticism.