Aaron Levie Warns 'Too Much Information' Can Derail AI Agents, Calls Data Chaos Key Enterprise Risk

Box Inc. (NYSE:BOX) Aaron Levie warned that the effectiveness of AI agents depends heavily on the quality and structure of company data, saying most enterprise failures stem from disorganized information rather than the technology itself.

Data Chaos Undermines AI Agent Performance

On Monday, in a post on X, Levie said "the biggest challenge" companies face in deploying AI agents is ensuring they operate with "the right constrained context" for each task.

He cautioned that "too much information or conflicting sources" can lead agents to draw incorrect conclusions, especially in environments where multiple systems of record exist.

Levie pointed to common enterprise issues, including outdated knowledge bases, duplicated documents and "tribal knowledge" systems where critical information is not formally documented.

He also noted the opposite risk, saying "too little information and the upside is highly limited" for AI agents, reducing their usefulness.

He added that "AI strategies are actually data strategy challenges in disguise," emphasizing that companies must prioritize how information is structured, maintained and accessed if they want AI systems to perform reliably.

This is true of all agents, not just coding agents. Probably the biggest challenge that most companies run into in their agent strategy is getting agents the right constrained context to work with for a task.

Too much information or conflicting sources, and the agent can easily… https://t.co/ewmJuZe0ZC

— Aaron Levie (@levie) May 19, 2026

Big Tech Expands AI Agent Push Across Products

Earlier, Alphabet‘s Google (NASDAQ:GOOGL) (NASDAQ:GOOG) reportedly tested an internal AI agent called "Remy" inside its Gemini app to perform tasks across daily life and work, signaling a shift toward more autonomous assistants beyond simple chat tools.

Airbnb Inc. (NASDAQ:ABNB) said AI generated 60% of its new code in the first quarter, while CEO Brian Chesky noted one engineer could now do the work of 20.

The company also said its AI support agent resolved 40% of customer issues without human escalation.

OpenAI reportedly accelerated plans for an AI agent smartphone targeting mass production by 2027.

Analyst Ming-Chi Kuo said the device could feature advanced AI-focused hardware and reach about 30 million shipments between 2027 and 2028.

The developments highlight how major tech firms are increasingly embedding AI agents into software, services and future hardware products.

Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.

Photo courtesy: Shutterstock