May 2026 Summaries

In December 2025, a cyberattack targeted Mexican government systems, exploiting 20 known and unpatched CVEs to exfiltrate 150GB of sensitive data, affecting around 195 million identities. The attacker utilized advanced AI tools like Claude Code and GPT-4.1 to swiftly identify and exploit vulnerabilities, outpacing traditional defensive measures. This incident highlights the accelerating challenge of managing vulnerabilities, as the time between disclosure and exploitation has drastically shortened, with many CVEs now exploited within a day. The traditional vulnerability management approach is becoming obsolete as defenders must adapt to a rapidly changing threat landscape by implementing strategies such as proactively limiting exposure, leveraging AI for faster remediation, and focusing on detecting behavioral anomalies post-exploitation. The institutional response, including changes by NIST to its CVE enrichment process, reflects the growing complexity and volume of vulnerabilities, necessitating a shift towards more agile and automated defense mechanisms.

Managing unstructured documents like invoices can be challenging due to the varied formats from different vendors, but a lightweight application using Box AI Extract and Files SDK streamlines this process efficiently. This application, built in under 130 lines of TypeScript, classifies invoices by expense category using Box AI and organizes them into folders with Files SDK, an open-source, community-built unified storage API for JavaScript/TypeScript. Files SDK provides a consistent interface for file operations across different storage systems, while its integration with Box allows for advanced functionalities like AI-powered classification. The app categorizes invoices through Box AI, which reads documents server-side and returns structured data, ensuring clean and consistent organization by automatically creating and moving files into category-specific folders. Additionally, the application offers potential for further enhancements, such as ingesting files from other sources, applying metadata templates, sharing categorized folders with stakeholders, and syncing data downstream, demonstrating the versatility and extensibility of the Files SDK and Box AI combination.

The extensive use of manual document handling in the insurance industry significantly slows down operations, consuming up to 70% of the insurance claim cycle and costing large carriers over $100 million annually. This inefficiency stems from the need to process unstructured data, which constitutes 90% of enterprise data, including essential documents like loss runs, broker statements, and medical records. AI-powered claims processing offers a solution by automating the classification, extraction, and formatting of document data, allowing skilled adjusters to focus on more critical evaluations rather than mundane data entry tasks. Companies like Shelter Insurance have successfully implemented AI tools, such as Box, to streamline their document processing workflow, reducing payout times from weeks to hours and recovering substantial adjuster hours. Moreover, the integration of AI into systems like Guidewire ensures secure and compliant document handling, while maintaining the necessary human oversight for complex cases. AI automation not only improves efficiency and reduces costs but also enhances competitive advantage by enabling faster, more accurate decision-making in both claims and underwriting processes.

Claude Opus 4.8, Anthropic's latest model, demonstrates significant improvements over its predecessor, Opus 4.7, particularly in tasks involving report drafting, financial analysis, and public sector applications. Evaluated using a rigorous suite of real-world enterprise tasks, Opus 4.8 consistently outperforms Opus 4.7 by producing more accurate and complete outputs in areas such as industrial goods reporting, legal NDA reviews, and corporate lending analysis. Its strengths are most notable in high-accountability sectors like financial services and the public sector, where precision and reliability are crucial. The model's enhanced capability in identifying critical details across various tasks reduces the likelihood of oversight, thereby increasing confidence in its outputs. Opus 4.8's superior performance is attributed to its ability to consistently detect crucial elements that its predecessor missed, thereby offering more thorough and dependable analyses for enterprise users.

Box is redefining its approach to becoming an AI-first company by integrating AI into its workflows, focusing on how work is done rather than just adding AI features. The company is leveraging its Box Intelligent Content Management platform to streamline processes with products like Box AI Agents, Box Extract, Box Hubs, Box Apps, and Box Automate. These tools facilitate the transformation of unstructured data into operational information, enabling faster, more consistent workflows across business areas like legal, HR, and procurement. Box Automate orchestrates processes by integrating tasks such as metadata extraction, document routing, and e-signatures within a unified platform, removing the need for disparate tools and manual handoffs. This approach not only speeds up tasks, reducing time for report generation significantly, but also enhances consistency and auditability by creating a cohesive system of record. Ultimately, Box's goal is to embed AI into its foundational processes, allowing teams to focus on strategic decisions rather than manual operations.

In a rapidly evolving collaborative landscape, Cisco’s Collaboration Customer Success team has leveraged a strategic partnership between Box and Webex to maintain operational excellence at scale across global markets. This partnership has provided a secure and intelligent ecosystem, transforming knowledge management and innovation through AI. By utilizing Box AI, the team efficiently accesses stored knowledge, reducing manual labor and enabling data-driven decision-making. A centralized repository offers standardized knowledge transfer, ensuring contractors worldwide maintain consistent quality. AI-driven insights from Box have facilitated new workflows and CRM integrations, optimizing processes such as summarizing account statuses and standardizing success plans. As they integrate Box Hubs and explore AI's potential to analyze unstructured customer feedback, the team seeks to align development efforts with customer needs without extensive data consolidation. This framework, combining storage, intelligence, and communication, is designed to be scalable and ready for future AI-driven initiatives, ensuring customers receive timely and intelligent support.

The Flue Framework is a programmable TypeScript harness developed by Fred K. Schott, designed for building AI agents that effectively interact with enterprise content by integrating with the Box content management platform. This framework allows developers to control the runtime environment around AI models, facilitating the creation of enterprise content assistants that use Box as the source of truth. The Flue harness bundles agents into an HTTP server, enabling deployment and interaction with Box tools such as search, files, folders, and metadata. By utilizing user context, the framework allows for personalized responses based on locale, region, and other parameters, ensuring answers are grounded in the relevant content. The integration with Box MCP allows agents to connect and authenticate with Box's server, leveraging the content layer provided by Box while maintaining flexibility in request handling and response adaptation. The framework supports lightweight personalization and custom workflows, making it suitable for various industries that require specific contextual responses.

Box CEO Aaron Levie and AWS VP of Data and AI GTM Rahul Pathak discuss the requirements for successful enterprise AI, emphasizing the importance of integrating advanced AI models with enterprise data and existing workflows. Amazon Bedrock plays a crucial role in this strategy by allowing businesses to maintain stable APIs and choose from various models, enabling flexibility in a rapidly evolving AI landscape. Pathak highlights that enterprises should not wait for perfect data to begin AI initiatives, as innovation should occur alongside modernization efforts. The collaboration between AWS and Box aims to connect AI systems to valuable enterprise content, using AWS's infrastructure and Box's capabilities in unlocking unstructured data. This partnership seeks to address the challenge of linking powerful AI models to existing business processes, providing a practical solution to enhance productivity and operational efficiency. The conversation underscores the need for a strategic approach to AI adoption, focusing on business objectives, data integration, and the establishment of governance frameworks to ensure scalable and sustainable AI deployment.

Box CEO Aaron Levie emphasized at a Virtual Summit that the true impact of AI agents lies in their deployment across enterprise workflows, where they can automate tasks, increase efficiency, and transform business processes. Over the past year, AI agents have evolved from answering questions to executing complex tasks, potentially allowing every employee to access expertise at unprecedented speeds. However, challenges remain, particularly in managing unstructured data, which contains critical business context but is often fragmented across systems. This fragmentation poses risks and challenges for deploying AI at scale, as agents might access outdated or incorrect information. To address these issues, enterprises need a platform like Box's Intelligent Content Management, which securely manages, organizes, and connects enterprise content with AI agents, ensuring efficiency and security. The platform is designed to transform workflows by enabling seamless access to essential information while maintaining robust governance and security, thus paving the way for a future where AI agents significantly outnumber human counterparts in the workforce.

During the Box Virtual Summit, OpenAI's Dominic Grillo and Box CEO Aaron Levie discussed the transformative role of AI agents in enterprises and the critical importance of governance. Grillo highlighted a significant improvement in AI capabilities, emphasizing the emergence of two types of agents: the "AI super assistant," which acts as an extension of a user's identity, and the "workspace agent," which functions as an AI coworker managing tasks like reviewing legal agreements. These agents can operate headlessly through APIs and workflows, accessing vast amounts of unstructured enterprise data to provide context and enhance automation. However, Grillo stressed the challenges of providing secure access to sensitive data and ensuring agents understand business context to deliver high-quality outputs. With AI agents becoming a persistent resource across applications, Levie and Grillo underscored the need for interoperability and robust governance strategies to manage identity, access, and data security, with Grillo assuring that solutions to these challenges are on the horizon.

AI agents are increasingly being integrated into enterprise operations, but their deployment comes with significant security challenges that outpace current protective measures. Bakhshi Malhotra from Box AI Security and Compliance and Heather Ceylan, Box's Chief Information Security Officer, emphasize the necessity of developing both deterministic and behavioral guardrails to mitigate risks like prompt injection and agent drift, where AI agents might deviate from intended actions due to overly permissive access. Unlike traditional assistant roles that merely provide answers, AI agents can autonomously trigger actions, leading to potentially severe consequences if misused. Organizations are urged to adopt a proactive security approach, focusing on risk-based governance and safe experimentation rather than hindering AI adoption. Box positions itself as a secure content layer that supports agent operations, advocating for a foundational approach to agent security instead of viewing it as an add-on. The conversation underscores the urgency of adapting security practices to keep pace with rapid AI advancements, while cautioning against the risks of shadow AI, where employees might deploy AI agents without IT or security oversight.

Box has released the results of its internal benchmarking, showing that the Gemini 3.5 Flash model significantly outperforms its predecessor, Gemini 3 Flash, in complex, agentic tasks across various industries. Gemini 3.5 Flash demonstrates a 74% accuracy rate compared to 62% for Gemini 3 Flash, with marked improvements in high-stakes domains like healthcare and life sciences. This improvement is attributed to the model's increased tool calls, thorough reading of source materials, and input verification before generating responses, enhancing its capability to handle intricate reasoning and data synthesis tasks. The introduction of Box MCP Server to the Gemini app expands the availability of Box's enterprise-grade content and agentic capabilities to a broader audience, including consumers, educators, and small businesses, emphasizing a shift toward AI models that can reliably interact with real-world data systems. This development highlights a step forward in making agentic workflows more viable for production use, as well as the collaboration between Box and Google to integrate these capabilities into widely used platforms.

Insperity, a Houston-based provider of HR solutions for small and mid-sized businesses, faced the challenge of modernizing its electronic records management system amidst growing volumes of sensitive HR data. To tackle this, the company partnered with Box to securely manage over 33 million HR records and streamline integration with Salesforce. By adopting Box's AI-powered tools, such as metadata extraction, Box Hubs, and Box Apps, Insperity enhanced compliance, improved document retrieval times, and optimized sales workflows. This transition not only enabled Insperity to handle increased data without proportional staffing increases but also facilitated faster onboarding and more efficient client service, with the company expecting significant time reductions in managing document requests. Box's secure content management and governance capabilities have become central to Insperity's enterprise content strategy, helping the company aim for 7-10% annual client-base growth while maintaining operational efficiency.

Box AI's structured extraction capabilities, featuring new struct and table field types, enhance the processing of complex procurement documents by allowing users to define schemas that closely match document structures. This innovation enables the transformation of supplier agreements stored in Box into structured outputs suitable for downstream systems, automation, and metadata management. Struct fields group related data into nested objects, such as vendor information, while table fields organize repeated entries like delivery schedules into structured arrays. The enhanced extraction agent further improves results by efficiently handling complex document layouts, making the extracted data ready for integration into workflows like ERP systems, CRM entries, or procurement platforms. These advancements reduce post-processing efforts and facilitate the seamless connection of unstructured data to actionable workflows, ultimately unlocking the potential of such data across various use cases.

Combining the AI capabilities of Claude with the contextual framework of Box creates a powerful tool for enterprises, enabling AI to operate effectively within specific business environments. While Claude provides the intelligence to draft documents, manage workflows, and automate tasks, Box supplies the necessary context by organizing and providing access to relevant information and maintaining compliance with existing permissions and processes. This partnership allows AI to transition from impressive demonstrations to practical, actionable solutions within a company, by ensuring that the model is not overwhelmed with irrelevant data and can focus on the most pertinent files. Box acts as the "work brain," enhancing Claude's functionality across various applications such as chat, cowork, code, and platform solutions, tailoring them to specific industry needs like financial services and legal workflows. The collaboration emphasizes the importance of integrating solid AI models with the right business context, retrieval methods, and execution strategies to truly transform enterprise operations.

The discussion on the Box AI-First Podcast, featuring Jesse Henning from Argonne National Laboratory, highlights how the integration of AI into mundane tasks can significantly boost productivity in scientific environments. At Argonne, AI is employed to streamline processes such as document review and travel expense management, allowing scientists to dedicate more time to research by quickly finding relevant information within extensive documents. Jesse emphasizes the importance of using AI to augment rather than replace human judgment, ensuring that sensitive and novel research findings remain protected from competitive exposure. Effective AI implementation requires structured content, governance, and a clear understanding of the problems it aims to solve, as demonstrated by Argonne's use cases. The conversation underscores that AI's value lies in its ability to reduce routine work and enhance human decision-making, suggesting a future where AI supports rather than supplants human expertise in professional settings.

Claude Managed Agents offer a streamlined solution for setting up AI agents by providing a managed runtime that handles key operational components, allowing users to focus on configuring the system prompt, tools, and servers. This service simplifies the process with an interactive Quickstart experience and predefined templates, enabling seamless integration with platforms like Box, Asana, Linear, and Databricks. Users need specific prerequisites, such as a Box account and developer token, and the process involves creating an agent, connecting necessary tools, setting up the environment, and testing prompts. An example use case is the Contract Clause Extraction Agent, which automates the extraction of metadata from contracts stored in Box and creates corresponding tasks in Asana or other platforms, turning contractual obligations into actionable tasks. The setup eliminates the need for custom infrastructure and coding, making it easier to implement intelligent workflows quickly.

Federal agencies are transitioning from AI pilot projects to integrating AI into mission-critical workflows, focusing on secure, governed content foundations that support compliance and human oversight. Events like AIScoop’s AITalks and the Box Federal Summit underline the importance of reducing content fragmentation and improving operational efficiency through AI, particularly in automating processes such as document intake and workflow acceleration. Box plays a significant role by helping agencies modernize content operations, emphasizing interoperability and auditability as essential for scaling AI responsibly. As AI becomes central to government modernization, challenges like content sprawl and outdated enterprise content management systems hinder progress, necessitating a shift towards unified, model-agnostic content platforms. This approach not only enhances security and compliance but also supports meaningful AI adoption, allowing agencies to focus on high-value use cases that improve decision-making and service delivery without replacing human oversight. The discussions at these events highlight that successful AI integration requires a human-centered approach, ensuring that technology complements public servants' efforts to fulfill their missions effectively.

Financial institutions are increasingly adopting AI and document automation to enhance efficiency, reduce costs, and improve client experiences. By leveraging technologies such as machine learning, optical character recognition (OCR), and robotic process automation (RPA), banks and investment firms can automate tasks like data capture, extraction, validation, and routing. This automation streamlines processes such as invoice approvals, loan origination, and customer onboarding, thereby reducing manual errors and ensuring compliance with regulatory standards. The trend is supported by significant investments in AI, which are expected to grow from $35 billion in 2023 to $97 billion by 2027, highlighting the industry's commitment to integrating advanced technologies into their workflows. Document automation not only speeds up approval cycles and reduces operational costs but also strengthens data security and compliance, crucial for the highly regulated financial sector. Companies like Box are providing platforms that enhance these capabilities with AI-powered automation, ensuring secure and efficient management of financial documents.

Ransomware continues to be a predominant threat in the realm of data breaches, frequently causing significant disruption to organizations by encrypting critical systems and demanding payment for their release. This method is favored by cybercriminals not only for its financial yield but also for its ability to inflict operational paralysis and reputational damage, compelling organizations to comply with demands. Both large enterprises and small-to-medium-sized businesses (SMBs) are vulnerable, with SMBs often lacking the robust defenses needed to mitigate such attacks. It is crucial for organizations to secure their own environments and carefully vet third-party platforms handling sensitive data, as reliance on these platforms can expose them to additional risks. The broader cybersecurity landscape remains perilous, with bad actors continually seeking to exploit any available sensitive information. As ransomware incidents persist, organizations must adopt comprehensive security measures and prepare for rapid recovery to minimize potential impacts.

Box has enhanced its capabilities by enabling the extraction and utilization of embedded metadata for digital asset management, making technical metadata searchable and available through the Box Apps dashboard. This metadata, which may include attributes like EXIF, ICC Profile, and JFIF, is represented in formats such as thumbnails, PDF, and text, and is generated on demand when files are uploaded. The process involves using the Box API to request file representations and employing metadata templates and scripts to map and apply these data points to images. This advancement allows users to display, search, and filter images based on both technical metadata and image content, facilitating more efficient management and retrieval of digital assets.

AI is revolutionizing the architecture, engineering, and construction (AEC) industry by transforming both the operational and information-management aspects, which are traditionally reliant on cumbersome, manual processes. It enhances efficiency in resource allocation, equipment maintenance, and job-site safety, while also addressing the industry's document-heavy nature by streamlining workflows and improving data accessibility. AI facilitates faster pre-bid work by validating bid packages and comparing subcontractor scopes, and it reduces compliance bottlenecks by automating data extraction and routing, thereby minimizing errors and ensuring up-to-date documentation. Companies like Novo Construction and Marx Okubo are leveraging AI to automate complex document processes, significantly reducing processing times and enhancing project management by creating centralized, searchable content repositories. This integration of AI in content management is driving a paradigm shift in the AEC industry, promising not only increased efficiency but also reduced risk and improved decision-making capabilities, ultimately fostering a more connected and agile construction environment.

Building an AI-driven document pipeline involves extracting structured data from unstructured content and making it actionable for team members, rather than leaving it as an ephemeral JSON blob. This architecture processes a Box folder containing vendor documents such as SOC 2 reports and contracts, using a multi-agent analysis to generate a risk report, which is then written back to Box with metadata and a review task for the security team. Box AI Extract Structured handles document intelligence, while LangChain Deep Agents manages the reasoning by deploying subagents in parallel to analyze different risk domains like security controls, compliance gaps, and contract risk. The output is synthesized into a final risk score, and findings are stored in a Markdown report shared within Box's collaboration environment. This setup allows for a separation of concerns, where the agent focuses on reasoning over structured data in a virtual filesystem, while Box handles file management, versioning, and collaboration, enabling efficient analysis without reinventing the wheel for document parsing and delivery systems.

The emergence of AI-driven technologies and the rapid evolution of the open-source ecosystem have introduced significant security challenges, as highlighted by recent incidents involving compromised AI tools and platforms. The time frame between discovering vulnerabilities and their exploitation has shrunk dramatically, putting immense pressure on the open-source supply chain. This has been exacerbated by AI-assisted tools that can identify vulnerabilities faster than human defenders can address them. Notably, a systemic architectural vulnerability in Anthropic's Model Context Protocol (MCP) SDKs, affecting numerous servers and millions of downloads, underscores the need for stronger security governance. The incident exposed the lack of effective vetting controls in AI-native dependency distribution channels, revealing a nascent but rapidly growing ecosystem prone to the same security failures witnessed in traditional software supply chains. The text emphasizes the urgency for robust governance models, including curation of marketplaces, capability-based permissions, sandboxed execution environments, and comprehensive signing and verification processes, to mitigate these risks. It also advocates for proactive engagement by security leaders to influence standards and governance early on, before these frameworks become entrenched. The overarching message is a call to action for the open-source community and enterprises to apply learned best practices to this new dependency ecosystem to prevent it from evolving into a chronic security crisis.

Anthropic has introduced new Claude solutions tailored for the legal industry, in collaboration with Box as a secure file system partner, to enhance AI-driven workflows while maintaining stringent security and compliance standards. These solutions address the "governance gap" by ensuring all AI activities respect existing Box permissions and policies, which mitigates risks associated with managing sensitive client data. The new tools transform Claude from a mere research assistant into an active agent capable of executing complex legal workflows, such as managing virtual data rooms, streamlining client intake and onboarding processes, and automating contract review for compliance. This integration aims to reduce manual work, lower costs, and improve client relationships by centralizing firm intelligence within Box, allowing law firms to operate more efficiently and securely. The solutions are compatible with existing tools like Anthropic Claude, OpenAI ChatGPT, Atlassian, and others, with plans for further integration with Salesforce Agentforce.

Artificial intelligence (AI) is revolutionizing business operations through workflow automation, which enhances efficiency by reducing manual intervention and enabling faster, more informed decision-making. By leveraging technologies such as machine learning, natural language processing, and intelligent document processing, AI-driven workflow automation processes information seamlessly across different systems, diminishing delays in tasks like document approvals and data integration. This approach replaces traditional rule-based systems, which often struggle with unstructured data, with more adaptable AI systems that can handle complex dependencies and execute tasks autonomously. Adopting AI in workflow processes not only improves data accuracy and scalability but also significantly reduces turnaround times and human error, allowing businesses to streamline operations and enhance productivity. Tools like Box Automate allow organizations to build AI-driven workflows that connect multiple AI agents for task execution, making it easier to manage and process large volumes of data efficiently. As AI technologies become more prevalent, businesses are increasingly integrating them to reshape work dynamics, striving for greater efficiency, accuracy, and decision-making capabilities.

Box faced significant challenges in maintaining real-time accuracy for folder size and item counts due to the high volume of file operations, leading to database overload and inefficiencies in their legacy system. To address this, Box transitioned from standard streaming patterns to a custom Dataflow-based solution using a hybrid session-gap approach, which effectively balanced database efficiency and user-visible updates. This new approach incorporated fixed-length session windows and a watermark-based mechanism, enabling more controlled and predictable aggregation while significantly reducing write traffic and infrastructure costs. The optimized pipeline achieved a 67% reduction in database load, saving approximately $300k annually and improving system reliability and user experience, demonstrating that strategic streaming aggregation can enhance both cost efficiency and performance without compromising user satisfaction.

In the examination of AI security challenges associated with the Model Context Protocol (MCP), new enhancements have been introduced to bolster governance and auditability, specifically through the implementation of Admin Controls for the Box MCP server. These controls aim to manage the interaction of AI agents with proprietary content by balancing automation of complex business processes with the associated risks of "write" actions, such as file movement or metadata updates. The new features include a centralized Admin Console that provides precision control, allowing IT administrators to customize agent capabilities according to their organization's risk tolerance and compliance requirements. This update enhances security by offering granular access tiers, custom tool configurations, invisible guardrails to prevent unauthorized actions, and full auditability for tracking configuration changes. Ultimately, these developments aim to integrate AI agents seamlessly with human employees while extending Box's security framework to accommodate advancing AI technologies.

Box has been recognized as a Leader in the 2026 Gartner Magic Quadrant for Document Management, highlighting its role in transforming document management into Intelligent Content Management. This approach involves securely managing enterprise content, making unstructured data actionable, and enhancing security frameworks with AI-driven workflows. Box aims to help organizations automate document-heavy processes and unlock information while maintaining compliance and governance. The platform supports scalability by integrating AI, workflows, and governance on a single platform, allowing teams to automate tasks and reuse workflows across departments without added complexity. With features like Box Extract, Box Shield Pro, Box Agents, and Box Automate, Box enables enterprises to capture metadata from documents, automate business processes, and create AI-driven workflows at scale. By providing tools and support for both technical and non-technical teams, Box empowers organizations to enhance speed, governance, and business outcomes, ultimately facilitating faster and more intelligent work processes.

The 2026 NAB Show in Las Vegas underscored the transformative impact of AI in the media and entertainment (M&E) industry, highlighting its evolution from supporting roles to becoming an integral part of operations by automating workflows in production, marketing, and distribution. AI in M&E is increasingly agentic, capable of understanding scripts and contracts, extracting key information, and triggering workflows, but its effectiveness hinges on its integration with content, which forms the core of every business operation. The show emphasized the importance of metadata as a critical component for AI to understand business context and automate processes, driving real ROI in the industry. As content creation accelerates, the need for scalable infrastructure and robust content management platforms like Box becomes vital to handle large content volumes without security risks, ensuring seamless workflow execution. The modern entertainment stack relies on the interplay between AI, metadata, automation, and security, with content as the central platform, enabling companies to move faster to market, scale creativity, and unlock greater value from their content.

Box Automate is an AI-native workflow solution that streamlines content-centric processes by integrating people, AI agents, and enterprise systems. It features a no-code, visual drag-and-drop builder allowing business teams to design workflows that respond to events, business logic, and AI-driven decisions. The workflows are initiated by flexible, event-driven triggers such as file actions, Box Sign events, and manual or scheduled executions. Once triggered, the Automation Engine orchestrates execution, evaluates conditions, and ensures reliable progression with support for conditional branching and dynamic variables. AI agents, including Box Agent and custom agents, can be utilized for tasks like content classification and data extraction, and workflows can pause for human oversight where necessary. Actions within Box and across connected systems are executed based on decisions, with capabilities for content and metadata management, notifications, and external API integrations. Box Automate provides a governed environment with centralized controls and compliance, enabling scalable, intelligent workflows that maintain security and visibility across processes.

Box Automate offers a transformative approach to enterprise workflows by leveraging AI agents to automate and orchestrate tasks around content, enabling organizations to redesign work processes with greater efficiency and context. By focusing on the accumulated burden of manual hand-offs and context reconstruction, Box Automate simplifies end-to-end workflows by using extracted metadata to trigger actions and route content effectively. Early adopters, such as Samsung and Argonne National Laboratory, have reported significant improvements in scalability and efficiency, highlighting the potential for AI-driven automation to reduce repetitive tasks and enhance decision-making. The solution integrates seamlessly with the Box content platform, ensuring content security and governance while enabling richer, more actionable automation. This approach shifts human roles towards more strategic tasks like judgment and exception handling, as AI agents manage metadata extraction, policy comparison, and workflow routing, ultimately allowing organizations to achieve more streamlined and scalable operations.

Microsoft has announced new capabilities for Microsoft Copilot Cowork and upcoming Box AI features, marking a significant evolution in AI technology that moves from simply answering questions to autonomously executing multi-step tasks. By integrating these AI agents with Box, enterprises can ensure that their work is powered by their most important and governed content. The integration is facilitated by the Box plugin, which connects Copilot Cowork to enterprise knowledge while maintaining security and eliminating the need for data migration. This development provides persistent security, allows enterprises to leverage AI without migrating data, and gives IT decision-makers confidence in data protection. With Box directly integrated into the Microsoft ecosystem, Copilot Cowork can act as a digital teammate, enhancing productivity across various applications like Outlook, PowerPoint, and Teams by executing complex tasks using information stored in Box. This integration positions Box as the file system for AI and is designed to work seamlessly with a range of tools, including those from Anthropic, OpenAI, Atlassian, and others, with further integrations planned for Salesforce Agentforce.

Guidewire and Box have teamed up to address the challenges faced by property and casualty (P&C) insurers in managing both structured and unstructured data. While Guidewire's platform efficiently handles structured data such as policy, billing, and claims information, Box focuses on managing the unstructured data that often remains unseen by analytics systems. This unstructured data includes documents, images, and correspondence crucial for understanding the context behind insurance metrics. Box provides a cloud-native repository and employs AI for dynamic insight extraction, automating workflows that traditionally required manual intervention. This collaboration enables insurance carriers to streamline processes, enhance decision-making, and accelerate operations by integrating structured financial data with the contextual richness of unstructured content, ultimately bridging the gap between data analysis and actionable insights.

Deploying AI agents within an enterprise requires secure and governed access to proprietary business content, facilitated by the Box Model Context Protocol (MCP) server. The Box MCP server enables AI agents to effectively interact with unstructured data while maintaining security, privacy, and governance, crucial for avoiding issues like context bloat, payload size limits, and data corruption. To address these challenges, strategies such as using structured text formats, leveraging MCP resources for content referencing, utilizing programmatic tool calling ("Code Mode"), and implementing signed URLs are recommended. However, these methods must also mitigate security risks, especially concerning prompt injection attacks, which can lead to data exfiltration. Box has implemented various measures such as admin controls, policy-based guardrails, and human-in-the-loop processes to enhance security. The ultimate goal is to enable AI agents to operate alongside human employees securely and efficiently, with the Box MCP server providing the necessary infrastructure for this integration.

Agentic workflows leverage AI agents to independently execute complex sequences of tasks, moving beyond traditional automation to adapt to real-time data and changing conditions. These workflows integrate with various platforms to streamline operations across departments like HR, marketing, and finance by automating data collection and analysis, thereby uncovering insights from unstructured business content. Utilizing large language models and tool integrations, AI agents can manage tasks such as document processing, error handling, and validation with minimal human intervention, ultimately enhancing decision-making speed and operational efficiency. The use of platforms like Box allows businesses to securely manage content while deploying agent-based workflows that improve scalability and data accuracy without increasing overhead. As organizations increasingly adopt these systems, the impact of agentic AI is expected to transform operations by providing more consistent and efficient processes that learn and adapt over time.

Box Markdown Editor streamlines the process of creating, editing, and managing Markdown documents by integrating these capabilities directly into Box, eliminating the need for multiple tools and reducing version confusion and content drift. This feature allows teams to draft, review, and publish documentation within the secure and governed environment of Box, making it ideal for technical and operations teams that rely on accurate and easily updatable content. The editor offers three editing modes—Code, Preview, and Split View—to accommodate different tasks, ensuring flexibility and precision during content creation. Furthermore, the integration with Box AI enhances the utility of Markdown files by making them queryable and useful as knowledge sources for AI agents, thereby connecting documentation to AI-powered workflows. Available to all Box customers on supported plans without additional configuration, Box Markdown Editor simplifies content management by keeping all documentation within the same collaborative space where teams already operate.

Organizations are struggling to scale AI initiatives beyond initial pilots due to data accessibility issues rather than technology limitations, as highlighted by Box CTO Ben Kus. Many companies mistakenly invest in advanced AI tools without first ensuring that their data is organized and accessible, often resulting in AI systems being unable to access critical business information trapped in legacy systems. Kus emphasizes the importance of starting with a solid data foundation, suggesting that organizations should first audit their data environments and establish a single source of truth with proper governance to prevent data leaks. He narrates a financial firm's journey from failure to success by initially focusing on simple data extraction and structuring before gradually scaling up to more complex AI functions. This approach of starting small with foundational capabilities and addressing specific bottlenecks has proven more effective than immediately pursuing complex AI implementations, ultimately leading to transformative business processes.