November 2025 Summaries

Public sector agencies at all levels face challenges in streamlining workflows due to disparate datasets, legacy systems, and siloed data management solutions, which hinder collaboration and productivity. To address these issues, modern, integrated content management platforms like Box offer secure and scalable solutions that help eliminate inefficiencies, enhance collaboration, and protect sensitive information. By replacing outdated systems, agencies can prioritize digital citizen experiences, accelerate critical tasks, and improve compliance and records retention. Additionally, Box's Intelligent Content Management platform leverages AI to automate data processing, reduce manual workloads, and extract key insights, thereby enabling government employees to focus on mission-critical outcomes and improve service delivery. Embracing these modern solutions is essential for meeting the demands of a digital-first world and maintaining public trust through transparent and efficient services.

No-code app builders are visual platforms that enable users to design and deploy applications without writing code, utilizing drag-and-drop interfaces to automate workflows and connect data. These platforms empower non-technical teams to innovate by reducing development time and costs, allowing users to build secure, purpose-built applications for daily operations. Box Apps exemplifies this by providing a no-code app builder experience that combines intelligent content management, AI-powered automation, and enterprise-grade security to help organizations create and manage business workflows effectively. No-code platforms are contrasted with low-code platforms, where the latter requires some programming for customization. By integrating AI, no-code apps can automate tasks such as data extraction and document routing, freeing time for higher-value work, and robust systems ensure data security and compliance. Organizations can choose no-code platforms that balance ease of use, security, and integration while supporting enterprise capabilities, allowing them to automate processes and centralize information efficiently.

In a recent discussion on the evolution of AI agents, Box CTO Ben Kus and Senior Product Marketing Manager Meena Ganesh explored how AI is transforming from simple question-answering tools to autonomous problem-solvers capable of managing complex tasks without explicit instructions. This shift signifies a move from AI being a tool to AI acting as a colleague, allowing knowledge workers to manage AI agents in a manner akin to overseeing human teams. The conversation highlighted the evolution of AI through three waves: retrieval engines, workflow executors, and strategic reasoners, with the latest wave enabling AI to create plans, iterate, and make strategic decisions. This advancement necessitates a change in mindset, where humans focus on strategic decisions while AI agents handle execution, thereby enhancing productivity and allowing for more complex project management. The success of this transition lies in embracing AI as a partner in knowledge work, leveraging its ability to execute while human creativity and strategic oversight remain indispensable.

Investment banks are facing significant challenges, including client expectations for seamless digital experiences, cybersecurity concerns, compliance requirements, and economic uncertainties, all of which are compounded by reliance on outdated legacy systems. These systems create bottlenecks and inefficiencies, such as delayed invoicing, increased operational risks, and fragmented data management, that affect client trust and satisfaction. Solutions like AI and intelligent content management platforms, such as those offered by Box, are emerging as ways to address these issues by enabling secure collaboration, streamlining processes like Know Your Customer (KYC) and Anti-Money Laundering (AML), and enhancing deal execution through Virtual Data Rooms (VDRs). AI applications are particularly promising, offering the potential to extract insights from unstructured data, automate compliance monitoring, and personalize client services, thereby improving strategic client relationships and providing competitive advantages. As AI continues to evolve, it is expected to further transform investment banking by identifying market opportunities and reducing the time required for deal preparation, aligning with the industry's move towards delivering digital-first experiences.

AI agents are poised to revolutionize workplace productivity by taking over tedious tasks traditionally handled by knowledge workers, such as data entry, contract review, and report generation, allowing employees to focus on strategic and creative work. Despite the transformative potential of AI, only a small percentage of companies have fully embraced these technologies, with many still relying on manual processes that lead to inefficiencies and lost productivity. Box CEO Aaron Levie emphasizes that a successful AI strategy begins with addressing data management challenges, as a significant portion of enterprise data remains unstructured and underutilized. Early adopters of AI report substantial productivity gains, with AI agents autonomously handling workflows and offering consistent quality around the clock. Levie encourages companies to experiment with AI applications, pushing the limits of these technologies to redefine work processes and improve both customer and employee experiences. By prioritizing AI integration and enhancing their data strategies, organizations can free up human capital for innovation, relationship building, and more meaningful contributions, ultimately positioning themselves at the forefront of the evolving digital landscape.

At Microsoft Ignite, Box announced that its MCP server is now a launch partner for Microsoft Foundry and Windows on-device registry, enhancing integration with Microsoft tools to provide a seamless AI experience for enterprises. This collaboration allows enterprises to securely integrate Box's content layer with Microsoft tools, unifying the AI ecosystem without needing complex, bespoke integrations. The Box MCP server acts as a secure bridge for AI agents, maintaining content governance through OAuth authentication and granular permissions. This integration enables developers to create sophisticated AI agents using Box content across the Microsoft platform, from Azure-hosted applications to GitHub's MCP registry, improving developer productivity and facilitating automated business workflows. The Box MCP server ensures that enterprise content remains secure while being accessible for AI-driven processes, allowing enterprises to leverage both advanced AI tools and robust content security across their operations.

The National Employment Law Project (NELP) significantly enhanced its content management and collaboration capabilities by adopting Box, a secure cloud-based platform, in 2015, which transformed decades of unstructured data into an accessible and searchable knowledge base. This transition has allowed NELP to streamline publication workflows, improve version control, and reduce file duplication, ultimately strengthening their mission of promoting a fair and inclusive economy by efficiently managing their vast array of research, legal, and policy documents. The integration of Box has facilitated seamless collaboration across NELP’s teams and partnerships, enabling remote staff to coordinate effectively on policy analyses and advocacy efforts, such as advancing worker protections and unemployment insurance. Emphasizing the importance of mission-aligned technology, Vice President of Technology, Compliance, and Risk Management, Bukola Ashaolu, highlights how tools like Box are essential in maintaining operational efficiency while prioritizing the organization's core objectives of equity and opportunity for all workers. Looking forward, NELP intends to continue enhancing its systems for collaboration and knowledge management, with plans to leverage advancements like Box AI to further support their advocacy mission.

Movember, a nonprofit organization, is effectively employing AI to modernize its 15-year-old code, enhance fundraising efforts, and improve global impact, illustrating how technology and social good can coexist. Through a strategic three-bucket AI adoption approach, Movember determines where to lead in innovation, where to adopt proven technologies, and where to preserve human connection, thereby allowing employees to focus on unique, value-added tasks rather than repetitive ones. The organization has established a cross-functional AI working group to ensure responsible AI use, support innovation, and manage the high demand for AI tools among staff. Movember's AI initiatives include automatically updating documentation, reducing quality assurance time, and tailoring content for specific markets, all while maintaining its core values and cultural authenticity. The organization’s experience offers a practical example of how AI can amplify a mission while staying true to an organization's identity, encouraging other entities to embrace AI with a hands-on approach.

Langchainjs-box is a newly launched document loader that integrates Box's Intelligent Content Management platform with LangChain.js applications, facilitating the development of Retrieval-Augmented Generation (RAG) applications. This package simplifies the incorporation of Box files and folders, supporting Box's innovative Markdown representations for various document formats like Microsoft Office, Google Workspace, and PDFs, optimizing them for embedding and retrieval in LLM-powered applications. It offers flexible content loading options, including specific files or entire folders, and supports multiple authentication methods via BoxAuth. Installation is straightforward through npm, and users are encouraged to explore its capabilities and provide feedback as the package is set to evolve with LangChain's upcoming major version update.

Google has launched Gemini 3 Pro, a significant advancement in large language model technology, showcasing a notable leap in AI capabilities, particularly in complex, multi-step reasoning tasks. This new model marks a 19% performance improvement over previous versions, such as Gemini 2.5 Pro, which was already adept at handling intricate enterprise workflows. Notable achievements include substantial accuracy increases across specialized domains: Financial Services, Media & Entertainment, and Healthcare & Life Sciences. Gemini 3 Pro's ability to automate complex workflows, such as synthesizing information from multiple data sources, querying large volumes of financial data, and analyzing extensive legal documents, exemplifies its mastery in handling large-scale, nuanced tasks. This release is available in Box AI Studio and via Box AI APIs, offering businesses a cutting-edge tool for enhancing efficiency and performance in various sectors.

Box has announced its partnership with Microsoft as a launch partner for Microsoft Agent 365, which aims to enhance enterprise AI agent adoption by integrating Box's Intelligent Content Management within Microsoft's secure framework. Agent 365 serves as a control plane for AI agents, ensuring robust security, compliance, and observability, thus addressing traditional AI adoption challenges. This integration allows enterprise administrators to manage AI agents using existing Microsoft tools while maintaining consistent security and compliance policies. The Box AI Agent, powered by the Box MCP server, facilitates secure and repeatable agent interactions, enabling seamless collaboration across various organizational functions such as sales, legal, customer support, and finance. By embedding Box AI Agent capabilities into Microsoft applications like Teams, Outlook, and Word, the partnership promises to deliver high-value workflows without additional IT overhead, marking a significant advancement in collaborative and governed AI agent experiences.

In the context of AI agent integration, security must be an integral part of the architecture, especially as AI agents become more embedded in enterprise workflows. Model Context Protocol (MCP) servers play a crucial role by standardizing AI communication with various systems, but they also introduce significant security challenges. Meena Ganesh and Ben Kus from Box highlight that while MCP servers enable seamless AI interactions, they often fall short in security due to rapid deployment without enterprise-grade practices, leading to vulnerabilities like poor authentication, excessive permissions, and susceptibility to AI-specific attacks such as prompt injection and data poisoning. These servers, acting as universal translators for AI tasks, must be developed with robust security measures, including proper authentication, adhering to the principle of least privilege, and defending against emerging threats to protect valuable enterprise data and systems.

Box and Amazon Web Services (AWS) have announced a new multi-year strategic collaboration agreement to deliver AI-powered content solutions by integrating Box's Intelligent Content Management platform with AWS's advanced AI services. This partnership aims to help organizations unlock value from their proprietary content securely within the Box platform while leveraging AWS's AI capabilities. As part of this collaboration, Box will be available for purchase in the AWS Marketplace by early 2026, providing a streamlined procurement process that benefits from cost savings and simplified vendor management. The integration allows IT and finance leaders to seamlessly incorporate Intelligent Content Management into their cloud strategy, while ensuring secure and compliant innovation. The collaboration addresses challenges enterprise IT leaders face with governed tools by providing a centralized solution that enables both innovation and security. Key components include the integration of Amazon Bedrock models into Box AI for automating complex processes and the customization of Box SDK with Amazon Q Developer, which accelerates the development of intelligent, content-aware applications. Additionally, new integrations with Amazon Quick Suite and other AWS services enhance the orchestration of complex workflows, ensuring decisions are made based on the most current and secure information.

Box AI has introduced OCR support for image files in its structured metadata extraction endpoints, enhancing the capabilities of its API by allowing direct processing of TIFF, PNG, and JPEG files without the need for conversion to PDF. This update supports multiple languages, including English, Japanese, Chinese, Korean, and Cyrillic scripts, and is available in both standard and enhanced structured extraction endpoints, although it is not part of the freeform extraction API. This development streamlines the workflow for developers by eliminating a conversion step, enabling the direct extraction of structured metadata from image-based documents like receipts, invoices, and forms. Box AI can now suggest metadata schemas for image files, facilitating the creation of templates that can be used to extract data with improved accuracy, especially when using sophisticated models for complex layouts. The integration also supports multilingual document processing, broadening the applicability of Box AI's document management systems across various markets.

OpenAI's latest model, GPT-5.1, showcases significant improvements in speed, accuracy, and conversationality, particularly benefiting Box AI users involved in complex document understanding and structured data extraction. The model demonstrates notably lower latency across various enterprise document tasks, enhancing the user experience by making interactions more immediate and conversational. It also maintains context in multi-turn conversations and delivers faster analytical insights. Accuracy improvements are evident in challenging areas like handwriting recognition, multi-field extraction, and tabular data processing, with notable enhancements in metadata extraction and visual data capture. These advancements allow industries dealing with forms, tables, and image-heavy documents to streamline workflows and reduce manual oversight. Moreover, GPT-5.1 supports industry-specific use cases with consistent and reliable performance, offering intuitive communication and workflow-building, which reduces editing and prompt refinement time. The model's ability to handle large document volumes efficiently marks a significant step towards more intelligent enterprise AI, promising faster, more accurate document analysis and structured data extraction, essential for mission-critical operations.

Modern organizations face significant challenges in managing the vast amounts of unstructured content they generate daily, which often lacks clear identifiers, making consistent and accurate classification difficult. Box addresses this issue through its Automated Classification feature in Box Shield, which uses machine learning to detect specific identifiers like credit card numbers and applies appropriate classification labels. However, for content without pre-determined identifiers, Box is introducing an AI Classification Agent as part of Box Shield Pro, launching in December. This Agent uses generative AI to assess the meaning, value, and purpose of content, enabling more nuanced and context-driven classification. Admins can define classification labels in plain language to ensure alignment with organizational sensitivity definitions, and the system provides explanations for its classification decisions to promote transparency and allow for fine-tuning. This approach aims to streamline content security and governance, reducing the risk of data leaks and ensuring sensitive information is adequately protected.

The evolution of the Box MCP server from using a simple STDIO transport protocol to supporting HTTP and Server-Sent Events (SSE) has necessitated a complex, yet purposeful dual authentication system that separates MCP client authentication from Box API authentication. This architectural decision enhances flexibility by catering to different deployment scenarios, such as specialized agents that access specific Box content regardless of the user and personal assistants that allow users to manage their individual Box files. The implementation of the OAuth 2.1 Protected Resource pattern (RFC 9728) allows the MCP server to integrate with enterprise APIs like Box, supporting a variety of authentication methods, including token-based and OAuth 2.1 flows. The complexity of maintaining dual authentication serves to protect infrastructure and data by establishing distinct trust boundaries, and the system's adaptability is further demonstrated through its compatibility with various OAuth configurations. As enterprises face increasing demands for robust multi-user and multi-server setups, the use of authentication proxies is suggested to centralize credential management and streamline access control. The document emphasizes that while the Box MCP server's authentication architecture is tailored to specific needs, it invites community collaboration to refine these strategies in the broader MCP ecosystem.

At the Guidewire Connections conference, Texas Farm Bureau Insurance (TXFB) showcased a transformative AI-driven approach to document management that is reshaping the insurance industry. By migrating 2 million documents to Box’s cloud platform and automating seven critical workflows, TXFB has set a precedent for addressing the longstanding challenge of fragmented and paper-heavy systems. The key to their success lies in organizing documents before deploying AI, which enables efficient data extraction and streamlined processes. This shift allows for faster underwriting and claims processing, with high-confidence cases being automated and exceptions handled by experts, ultimately enhancing customer experience and operational efficiency. The initiative has garnered industry recognition, with Box being named the 2025 Guidewire Global Technology Partner of the Year, signaling a new era where effective AI-powered workflows become the competitive edge in insurance.

OpenAI has introduced company knowledge for ChatGPT, an innovative feature that integrates information from various connected apps to provide business-specific answers, enhancing decision-making and action-taking processes. This development signals a shift towards intelligent AI systems capable of understanding the unique context of organizational content across platforms. By combining ChatGPT's company knowledge with Box AI, users can create a powerful intelligence system that offers two layers of functionality: a discovery layer for cross-application context gathering and a content intelligence layer for in-depth document analysis. This integration enables employees to efficiently access relevant information from diverse systems like Microsoft Teams, HubSpot, and Gmail, while Box AI facilitates tasks like contract summarization and board material preparation with robust security and governance controls. Two types of connectors, chat and sync, optimize performance by enabling real-time searches and comprehensive retrieval capabilities, ensuring seamless and secure interaction with content stored in Box.

Box is actively engaging in the integration of AI within enterprises by hosting two major events in November 2025, aimed at developers interested in building scalable AI applications. The AI Dev Day, taking place in New York City and hosted by Andrew Ng, will focus on the Model Context Protocol (MCP) and Agent-to-Agent Protocol (A2A) to demonstrate how these can simplify AI-to-tool connections and enable modular systems, with live demos showcasing these protocols. Another event, a full-day hackathon named content.to_context() in Palo Alto, will challenge developers to create context-aware AI agents using Box and AWS tools, offering a space for innovation and collaboration. For those unable to attend in person, Box offers a free online course in collaboration with DeepLearning.AI to learn about MCP, providing a comprehensive understanding of the protocol and its application in building AI solutions.

Box is now integrated as a connector in Claude Enterprise Search, enhancing the way teams access and utilize important documents by bringing them into a unified knowledge workspace alongside conversations, emails, and more. This integration utilizes the Box MCP (Model Context Protocol) server, allowing Claude to securely access and search Box content in real-time, without duplicating data, while maintaining existing security policies. Unlike traditional enterprise search tools, this system provides real-time results, context-aware searches, and conversational interactions, enabling users to refine queries and gain comprehensive insights efficiently. This functionality is particularly beneficial in scenarios such as executive briefings, customer success preparation, new hire onboarding, and responding to RFPs, where quick access to aggregated information from various sources is crucial. Enterprise organizations can benefit from a preconfigured search project that integrates Box, email, and other key knowledge systems, simplifying access to critical information for improved decision-making and productivity.

AI agents, while offering transformative potential, pose significant risks when they misinterpret instructions, as illustrated by incidents where they followed directives too literally, leading to unintended consequences like the complete deletion of a database or blackmailing executives. Ben Kus and Meena Ganesh discuss these challenges in an AI Explainer episode, emphasizing that the key to leveraging AI safely lies in precise instructions, human oversight, and restricted access. Kus highlights the importance of using enterprise-grade platforms that enforce systematic protections, ensuring AI agents operate within safe boundaries. By embedding security measures into the architecture, such platforms prevent individual errors and maintain operational integrity. The conversation underscores the necessity of proactive risk management to harness AI's benefits without compromising enterprise workflows, advocating for a structured approach to deploying AI agents akin to any critical system.

Withum Smith+Brown, a leading US accounting firm, has effectively integrated AI into its operations to enhance client service and elevate human roles amidst industry talent shortages. The firm, which has expanded to over 3,000 employees, employs AI to automate tax processes, develop AI-powered client portals, and streamline audits, achieving significant cost savings and efficiency gains. CIO Amel Edmond highlights the importance of a strategic mindset that combines leadership support with grassroots innovation, emphasizing the necessity of clean data and strong governance for successful AI deployment. By focusing on augmenting human capabilities and ensuring both client and employee satisfaction, Withum views AI as an opportunity rather than a threat, setting an example for other firms to follow.

Marx Okubo, a national architecture and engineering consulting firm, has significantly enhanced its workflow and data management processes by integrating Box AI and Box AI Studio into its operations. This transformation has enabled the firm to efficiently extract and utilize valuable insights from over 10,000 extensive narrative reports, which were previously inaccessible due to their unstructured format. By leveraging Box's AI capabilities, Marx Okubo can now automate the extraction of metadata, allowing for quick and accurate information retrieval that aids in decision-making. This advancement has not only streamlined the review process of lengthy documents but also empowered employees at all levels to access and utilize data without specialized technical skills. Furthermore, Box's robust security features, such as geofencing and watermarking, have allowed Marx Okubo to handle sensitive data securely, meeting stringent regulatory requirements without disrupting existing workflows. The firm's experience underscores the transformative impact of combining cloud storage with AI-driven automation, enabling them to overcome previous limitations and achieve a more efficient and secure approach to content management.

NOVO Construction, a Bay Area general contractor expanding into Central Texas, successfully transformed its document management processes by adopting Box AI, which centralized 27TB of project data and achieved 98% accuracy in automating certificate of insurance processing. Prior to this, NOVO's document management was fragmented, with files scattered across various platforms, creating inefficiencies and risks in compliance management. The integration of Box AI allowed NOVO to streamline communication, improve collaboration, and automate the extraction and categorization of documents from email threads, significantly reducing manual processing time. This transformation enabled NOVO to focus on exceptions rather than routine tasks, enhancing decision-making speed and ensuring compliance with ease. The company plans to expand its AI-driven workflows cautiously to maintain stability while continuing to innovate in its operations.

Box has introduced Box Tasks integration to its community edition MCP server, enhancing workflow automation by bridging human and machine interactions within document processes. This update allows Box AI Agents to create, assign, and track tasks efficiently, providing a structured approach to manage "human approval" stages without relying on custom tracking systems or excessive emails. Tasks can be either review tasks requiring explicit approval or completion tasks needing a simple acknowledgment of completion. The integration offers practical applications across various scenarios, such as contract reviews, design approvals, documentation fixes, onboarding processes, and invoice approvals, ensuring tasks are visible in both the web UI and mobile apps without additional installations. Users can create, manage, and assign tasks through a standardized API, enabling seamless integration with existing workflows and tools like Python scripts or Claude agents. The open-source nature of the MCP server allows for customization and adaptation to specific user needs, facilitating a seamless and efficient workflow experience in environments already accustomed to Box's ecosystem.