Uploaded July 2026 | Updated September 2026, 3 weeks ago
*Featured in this video:* Ben Kus, Chief Technology Officer at Box
*Executive summary:* Box partnered with Google Cloud to power Box AI, a new suite of AI agents, integrated directly into its leading content management and workflow platform. Built with customer choice of leading AI models, including Google’s Gemini models, Vertex AI, and BigQuery, Box AI brings the power of AI agents to uncover insights, find and structure information, extract relevant context, generate new content, and automate multi-step tasks — all underpinned by Google Cloud’s foundational scale, enterprise-grade capabilities, and security. Now, Box’s customers are saving countless hours in their critical business processes, from loan processing to legal discovery, with the confidence they can harness the untapped value for enterprise content, safely and securely.
*Challenge:* As much as 90% of today’s enterprise data is unstructured, meaning employees must spend hours manually searching through, analyzing, and understanding information to find answers and gain insights to fuel their everyday workflows. Box wanted to build AI that could help customers get the most from their valuable content, while still meeting its enterprise-grade standards for security, compliance, and privacy.
*Solution:* Building on its partnership with Google Cloud to automate content processing and metadata extraction, Box used Google’s Gemini models, Gemini Enterprise Agent Platform (formerly Vertex AI), and BigQuery to help power Box AI: A suite of pre-built and customizable AI agents that not only rapidly process and synthesize unstructured information but can also proactively take action on it. Gemini’s advanced reasoning, multimodal understanding, and long-context performance — supporting up to 2 million tokens — enables Box AI to quickly find information, surface critical insights, extract key context, classify documents, and even automate complex, multi-step workflows. In addition, Google Cloud helps Box build trusted, governed, and permissions-aware AI, providing the foundational scale, enterprise-grade capabilities, and security that can meet the requirements of even the most highly regulated industries.
*Results:* With Google Cloud, Box AI is transforming unstructured information into actionable data, safely and securely. Customers are already working smarter, saving countless hours in their daily workflows — from loan processing to legal discovery — to unlock the full potential of their enterprise content.
*Key takeaways and highlights from our interview with Ben Kus, I'm Chief Technology Officer at Box:*
→ “Box AI is everything that Box does for AI, bringing the power of AI agents to your unstructured data. We provide agents that find information, do research, then we also give you the ability to fully customize these. And you can say, here’s an objective and here’s some things I want you to do, and then have the agents make a multi-step plan and execute that.”
→ “We use Gemini to power our Box AI, which is one of the best models in terms of being able to solve the hardest problems. It’s also one of the fastest. For example, one of our customers, they were doing an audit, and they needed to go through all of their data to see which ones were customer related data or not. They were able to take AI agents and have them do that work, and within a day, they had it actually solved.”
→ “We use BigQuery quite extensively for processing the structured aspect of the unstructured data, and then, of course, we use Vertex AI and Gemini. Two million tokens on some models will take many minutes or an hour just to process. With Gemini, it can fly through that much faster, which then makes it actually feasible to use in many use cases.”
→ “Google makes up a significant part of our infrastructure and to provide a foundation of scale, a foundation of enterprise-grade, a foundation of security – that’s important for our customers and for us, as we at Box build solutions that are targeted for them.”
*Google Cloud products used:* Gemini, Gemini Enterprise Agent Platform (formerly Vertex AI), BigQuery
*Learn more:*
→ From content to action: Box and Google Cloud introduce AI agents for the enterprise blog.box.com/content-action-box-and-google-cloud-introduce-ai-agents-enterprise
→ Google Cloud Next 2025: How Box and Google Cloud are transforming enterprise content with AI blog.box.com/google-cloud-next-2025-how-box-and-google-cloud-are-transforming-enterprise-content-ai
→ Box transforms metadata extraction with Gemini cloud.google.com/customers/box?hl=en
*Featured in this video:* Ben Kus, Chief Technology Officer at Box
*Executive summary:* Box partnered with Google Cloud to power Box AI, a new suite of AI agents, integrated directly into its leading content management and workflow platform. Built with customer choice of leading AI models, including Google’s Gemini models, Vertex AI, and BigQuery, Box AI brings the power of AI agents to uncover insights, find and structure information, extract relevant context, generate new content, and automate multi-step tasks — all underpinned by Google Cloud’s foundational scale, enterprise-grade capabilities, and security. Now, Box’s customers are saving countless hours in their critical business processes, from loan processing to legal discovery, with the confidence they can harness the untapped value for enterprise content, safely and securely.
*Challenge:* As much as 90% of today’s enterprise data is unstructured, meaning employees must spend hours manually searching through, analyzing, and understanding information to find answers and gain insights to fuel their everyday workflows. Box wanted to build AI that could help customers get the most from their valuable content, while still meeting its enterprise-grade standards for security, compliance, and privacy.
*Solution:* Building on its partnership with Google Cloud to automate content processing and metadata extraction, Box used Google’s Gemini models, Gemini Enterprise Agent Platform (formerly Vertex AI), and BigQuery to help power Box AI: A suite of pre-built and customizable AI agents that not only rapidly process and synthesize unstructured information but can also proactively take action on it. Gemini’s advanced reasoning, multimodal understanding, and long-context performance — supporting up to 2 million tokens — enables Box AI to quickly find information, surface critical insights, extract key context, classify documents, and even automate complex, multi-step workflows. In addition, Google Cloud helps Box build trusted, governed, and permissions-aware AI, providing the foundational scale, enterprise-grade capabilities, and security that can meet the requirements of even the most highly regulated industries.
*Results:* With Google Cloud, Box AI is transforming unstructured information into actionable data, safely and securely. Customers are already working smarter, saving countless hours in their daily workflows — from loan processing to legal discovery — to unlock the full potential of their enterprise content.
*Key takeaways and highlights from our interview with Ben Kus, I'm Chief Technology Officer at Box:*
→ “Box AI is everything that Box does for AI, bringing the power of AI agents to your unstructured data. We provide agents that find information, do research, then we also give you the ability to fully customize these. And you can say, here’s an objective and here’s some things I want you to do, and then have the agents make a multi-step plan and execute that.”
→ “We use Gemini to power our Box AI, which is one of the best models in terms of being able to solve the hardest problems. It’s also one of the fastest. For example, one of our customers, they were doing an audit, and they needed to go through all of their data to see which ones were customer related data or not. They were able to take AI agents and have them do that work, and within a day, they had it actually solved.”
→ “We use BigQuery quite extensively for processing the structured aspect of the unstructured data, and then, of course, we use Vertex AI and Gemini. Two million tokens on some models will take many minutes or an hour just to process. With Gemini, it can fly through that much faster, which then makes it actually feasible to use in many use cases.”
→ “Google makes up a significant part of our infrastructure and to provide a foundation of scale, a foundation of enterprise-grade, a foundation of security – that’s important for our customers and for us, as we at Box build solutions that are targeted for them.”
*Google Cloud products used:* Gemini, Gemini Enterprise Agent Platform (formerly Vertex AI), BigQuery
*Learn more:*
→ From content to action: Box and Google Cloud introduce AI agents for the enterprise blog.box.com/content-action-box-and-google-cloud-introduce-ai-agents-enterprise
→ Google Cloud Next 2025: How Box and Google Cloud are transforming enterprise content with AI blog.box.com/google-cloud-next-2025-how-box-and-google-cloud-are-transforming-enterprise-content-ai
→ Box transforms metadata extraction with Gemini cloud.google.com/customers/box?hl=en
