TLDR: Box has rolled out a new suite of AI agents designed to revolutionize enterprise content management by offering instant risk assessment, deep research, and streamlined content creation. These agents leverage leading AI models to automate complex tasks, accelerate decision-making, and enhance productivity within a secure, permissions-aware environment.
Box, a prominent intelligent content management (ICM) platform, is significantly enhancing its offerings with the introduction of advanced AI agents aimed at transforming how businesses interact with their vast amounts of unstructured data. These new AI-powered tools are specifically engineered to provide instant risk assessment, conduct deep research, and facilitate intelligent content workflows, marking a pivotal step in enterprise AI adoption.
The core of this innovation lies in Box AI Agents, which enable users to engage directly with their content, including critical documents like Statements of Work (SOWs), leases, and thousands of other files. Users can now instantly ask questions, summarize complex terms, compare clauses, and, crucially, ‘surface risks’ embedded within their content. This capability is powered by integration with top foundation models from industry leaders such as OpenAI, Anthropic, and Google.
Box AI Agents operate on a dynamic agentic reasoning framework. This framework guides the agents through a sophisticated process: first, understanding the user’s objective (Plan); then, securely discovering relevant content within the Box platform (Discover); intelligently analyzing the information to extract insights (Reason); and finally, taking action or generating an output (Act). This entire process is meticulously designed to adhere to stringent organizational controls and leverage central data, ensuring both efficiency and data governance.
Beyond risk assessment, the new agents offer a range of breakthrough capabilities:
Smart, Efficient Searching: The AI search agent facilitates quick lookups, rapidly surfacing targeted answers, such as an expiration date in a contract or a client’s name in meeting notes. It also handles more complex queries by employing high-precision semantic analysis to uncover hidden relationships and insights across large volumes of documents.
Deep Research: These agents accelerate the insight generation process through AI-driven analyses across extensive content libraries, automatically identifying key trends and connections without requiring manual review.
Box Extract Agents: Designed for precision, these agents instantly pull key details from various content types, including contracts, forms, and images. They transform unstructured data into structured metadata, which can be customized with defined fields and extraction logic.
Content Composition: The agents can quickly generate marketing copy, product briefs, RFP responses, and executive summaries. This content is accurate, context-aware, and grounded in the user’s files, eliminating the need to switch between different tools.
Box Automate: A new agentic workflow automation solution, Box Automate, enables organizations to orchestrate work seamlessly across different agents and human teams. This streamlines processes, accelerates decision-making, and automates tasks that were previously manual.
Enhanced Box Apps: Box’s no-code solution for building intelligent applications now includes powerful AI capabilities, allowing organizations to create dynamic dashboards that integrate content, metadata, users, and workflows with agent-assisted analysis.
Security and compliance remain paramount for Box. The company emphasizes enterprise-grade security, compliance, and governance, ensuring that AI agents operate with permissions-aware models and granular access controls. This commitment extends to Box AI Studio, where administrators can restrict agent access to only authorized users, ensuring secure AI deployment.
Recent developments underscore Box’s aggressive push into AI. In September 2025, Box announced its next-generation AI Agents, including Box Extract and Box Automate, alongside expanded AI capabilities in Box Apps. At BoxWorks 2025, held in September, the company showcased how AI agents are already delivering tangible benefits, with examples of recovering 30 hours of manual work per week and scaling client onboarding throughput by nine times. Furthermore, an AI update on October 17, 2025, noted that Box is already deploying ‘Skills’ with Anthropic’s Claude.ai, Claude Code, the API, and the Agent SDK, enhancing Claude’s consistency and contextual awareness for professional workflows. Earlier in May 2025, Box expanded its AI agents for search, deep research, and data extraction, and introduced a dedicated Box AI Agent for Microsoft 365 Copilot, extending its content management capabilities into Microsoft applications.
Also Read:
- Microsoft Enhances 365 Copilot with Anthropic’s Claude AI Integration
- Anthropic Unveils ‘Skills for Claude,’ Empowering AI Agents with Enhanced Task Capabilities
These advancements position Box at the forefront of leveraging AI to create more intelligent, secure, and efficient content management solutions for enterprises worldwide.
