January 2026 Summaries

In 2025, the AI landscape matured significantly, evolving from basic chatbots to sophisticated systems that extract and classify data, enhance internal processes, and ensure security, particularly within the public sector. As organizations plan for 2026, AI is expected to become a foundational utility akin to electricity, integrating deeply into mission-specific functions across government agencies. Discussions with Box executives, Murtaza Masood and Jason Gray, highlight the shift towards using AI in existing platforms to improve efficiency and service delivery without additional budgets, focusing on use cases that yield measurable ROI, such as self-service AI agents for public information access and streamlined internal processes. Emphasis is placed on choosing FedRAMP High-authorized platforms that consolidate multiple functions, ensuring security and compliance, while AI deployments should prioritize governance, reducing risk and complexity. Agencies are advised to align AI initiatives with strategic objectives, demonstrating tangible ROI and responsibly managing taxpayer funds, which can drive future modernization support.

Bluedog Design, a leading product-innovation consultancy for consumer packaged goods, has revolutionized its approach to managing vast digital archives through a partnership with Box AI, transforming disorganized, hard-to-search files into a dynamic, AI-powered digital asset management system. This collaboration allows Bluedog to extract rich metadata from decades of product photography, enhancing searchability and providing their Fortune 100 clients with a comprehensive global view of packaging trends and market evolution. The integration of Box AI aligns with Bluedog’s values-first approach, emphasizing security and collaboration without compromising identity management, and supports a modular technology stack that includes tools like Slack, Microsoft 365, and Adobe Creative Suite. Box’s AI capabilities have enabled Bluedog to efficiently manage content while upholding its commitment to a people-over-profits ethos, allowing for seamless innovation and upskilling opportunities. This strategic adoption of AI demonstrates that even small teams can achieve significant digital transformation by prioritizing security, values, and human-centric technology solutions, ultimately proving that technology should enhance the enjoyment and effectiveness of work.

Content classification is fundamental to effective content security as it enables organizations to understand and manage the sensitivity of their data, ensuring appropriate protection measures are applied. Box Shield Pro enhances content security by introducing automated classification tools that leverage AI and rule-based systems to apply labels such as Public, Internal, Confidential, and Restricted, which dictate access, sharing, and retention policies. The classification process helps prevent human error and ensures compliance by providing clear, enforceable labels that guide security measures across an organization's content lifecycle. By automating classification and integrating it into Box Shield, organizations can protect sensitive information efficiently, reduce the burden on manual processes, and maintain security without hindering business operations. As Box continues to innovate, these classification-driven security measures expand to cover more nuanced content and provide advanced access controls, reinforcing the importance of classification as the central element in content protection strategies.

World Kinect, a global energy management company, has dramatically expanded its data volume and user base over the past decade, prompting the adoption of Box Shield Pro to enhance content security and collaboration. With the transition from manual to AI-powered content classification, World Kinect can now manage its vast 240 terabytes of data more efficiently, allowing its IT team to focus on strategic security initiatives rather than time-consuming manual classification. The introduction of AI classification helps address the shortcomings of traditional automation by understanding the context and purpose of content, turning security policies written in natural language into actionable classification rules. This shift towards intelligent content management not only reduces the risk of shadow vulnerabilities but also transforms AI into an auditable process, crucial for regulated industries, and enables the IT team to engage in high-value work. The experience highlights the importance of understanding data usage before implementing AI solutions, illustrating a broader shift in enterprise security from whether to automate content classification to how to do it effectively, combining human insight with AI scale.

Box has introduced a confidence score feature to its AI-powered metadata extraction service, providing users with a probabilistic measure of extraction accuracy for each field. These scores, ranging from 0 to 1, indicate the likelihood of an extracted field being correct, aiding users in determining which extractions need human verification. The scores are generated through consistency analysis of responses from the language model, with high scores reflecting consistency across different prompts. Users can incorporate confidence scores into their workflows by including a specific parameter in their API requests, enabling them to programmatically route lower-confidence fields for manual review. While high-confidence scores suggest reliable extractions, they are not guarantees, and critical data should be cross-verified. The feature is currently supported by Google Gemini models and is limited to the /ai/extract_structured endpoint. Confidence scores are designed to optimize extraction workflows by prioritizing human oversight where necessary, making them particularly useful in high-volume scenarios where manual review of every field is impractical.

In collaboration with Anthropic, the introduction of MCP Apps within the Box Connector in Claude brings a significant update by integrating visual interfaces into AI interactions, transforming enterprise content engagement. This update addresses the limitations of traditional Model Context Protocol (MCP) servers, which were previously confined to text and structured data, by introducing a dynamic that enhances the user experience through visual elements in chat interfaces. This improvement allows users to quickly identify and interact with high-value content such as diagrams and documents without needing to switch contexts, thereby maintaining focus on tasks. The Box Connector now offers tools for efficient content management, deep insights extraction, advanced search capabilities, and enhanced collaboration. It connects seamlessly with existing tools like Anthropic Claude, OpenAI Agent Builder, and Microsoft Copilot Studio, among others, and is poised to work with upcoming platforms like Salesforce Agentforce. This development is designed to enhance workflows by connecting unstructured data with leading AI models, ensuring that enterprise interactions are more intuitive and efficient.

Structured metadata in Box serves as a powerful tool for enhancing search, filters, retention policies, automation, and workflows, and with Box AI, users can extract structured data from unstructured documents and apply it as metadata in a systematic manner. This process is exemplified by building a simple Python CLI that orchestrates the extraction of specific fields from files stored in Box, transforming them into metadata using Box AI Extract Structured, and then writing them back to the files. Central to this implementation is the use of an "agents.md" file, which acts as a comprehensive specification for the project, detailing the expected CLI interface, authentication methods, and necessary SDK imports, thereby allowing AI coding tools to produce consistent and predictable results without needing extensive prompts. The architecture involves a straightforward workflow where the CLI facilitates the extraction, normalization, and metadata writing process, maintaining a clear separation of responsibilities within the code to ensure simplicity and ease of maintenance. This method not only enhances the reliability of AI-assisted development but also provides a robust framework that can be adapted for various workflows beyond metadata extraction, emphasizing the importance of defining project constraints and structure within the repository for effective collaboration with AI tools.

Box's Intelligent Prospecting Agent (IPA) has successfully moved from its Pilot phase, where it demonstrated significant time-saving and personalization capabilities, to the Rollout and Scaled Adoption stages, which are critical for integrating AI into business operations. During the Rollout phase, the focus is on ensuring the agent's reliability and accuracy, managing systems integration, and safeguarding data access, with the aim of making the AI agent operational for a broader user base. Scaled Adoption involves redesigning workflows to incorporate the agent into daily operations, ensuring it becomes a core tool rather than a side tool, and measuring its impact on productivity and revenue. The process requires both a functional leader to set expectations and an AI manager to facilitate user adoption through training and communication. As the agent evolves, it is important to maintain a clear understanding of its capabilities and limitations to foster trust and effective use. Ultimately, the success of the IPA and similar initiatives is determined by their ability to transform workflows and deliver measurable business outcomes.

Barnett Capital, a Chicago-based family office lender, has significantly enhanced its underwriting processes by integrating Box with Salesforce, thereby automating the extraction of key information from leases and appraisals. This integration has resulted in a 60-fold increase in underwriting speed, allowing the company to answer 35-40 questions per document in about a minute, compared to the manual process that previously took an hour. By centralizing documents on Box and utilizing Box Extract, Barnett has streamlined its document management, improved data security, and reduced the need for additional headcount despite business growth. Additionally, the use of Box Forms and Box Shield has enhanced the security and efficiency of client documentation processes, enabling the secure sharing of sensitive financial documents. Barnett's approach to leveraging technology for intelligent content management has not only improved its operational efficiency but also provides a roadmap for other companies aiming to modernize their processes while maintaining quality and control.

Adobe is addressing the challenge of scaling expertise across its organization by integrating AI to emulate its top experts and decision-makers, thus extending rather than replacing human expertise. This approach is part of a broader strategy to attract and retain top talent by providing access to innovative tools and fostering a culture where employees can experiment with cutting-edge technology. Adobe employs a structured "A through F" framework to guide safe AI implementation by considering key factors such as team involvement, technology used, and data input. This framework transforms data governance into an enabler rather than a barrier. Furthermore, Adobe utilizes a five-layered approach to AI pilots, which combines foundational AI models with subject-matter-specific and expert knowledge, thereby making expert insights accessible to the entire workforce. This strategy not only enhances productivity and efficiency but also ensures that specialized knowledge is shared throughout the organization.

Artificial intelligence in business, once seen as unreliable and over-hyped, is now gaining traction as AI systems become more reliable and effective in high-stakes environments. This shift is driven by the convergence of structured workflows with the flexibility of generative AI, allowing AI agents to complete tasks and collaborate efficiently. The key advancement lies in multi-agent collaboration, where specialized AI agents work together to streamline complex processes, such as responding to proposals, thereby saving time and enhancing productivity. Open standards like the Model Context Protocol (MCP) and the Agent2Agent Protocol (A2A) facilitate interoperability among different AI agents, enabling diverse systems to communicate and coordinate effectively. Trust and governance are crucial, requiring IT teams to maintain control over AI capabilities and ensure that these systems operate within defined boundaries. The evolution of AI is not about replacing human judgment but augmenting it by handling routine tasks, allowing humans to focus on creativity and strategic thinking. This AI transformation is marked by practical, incremental improvements rather than a single breakthrough, presenting significant opportunities for businesses that can effectively manage and integrate these technologies.

Deep Agents, a framework from LangChain, offers a structured approach to building complex systems by using orchestrators, sub-agents, and persistent memory to prevent the chaos often seen in multi-agent systems. In an implementation for auto loan underwriting, the orchestrator acts as a coordinator, delegating specific tasks such as document extraction, policy interpretation, and risk calculation to specialized sub-agents, each with limited and focused responsibilities. This separation of concerns ensures that each sub-agent is isolated and cannot access the full system, thus avoiding the "everything talks to everything" problem. The system supports a persistent audit trail, crucial for compliance in regulated industries, by recording every decision, calculation, and data extraction in a structured manner. Box AI is integrated to handle document intelligence, allowing for natural language queries and structured data extraction without the need for manual PDF parsing. This setup allows the orchestrator to manage workflows effectively and focus on business logic rather than infrastructure, providing a reliable and debuggable system.

Box Extract is a sophisticated AI-driven solution developed by Box to address the reliability challenges of enterprise AI systems, particularly in extracting consistent and trustworthy data from unstructured content like contracts and invoices. The system leverages a multi-faceted approach that includes model orchestration, document preparation, semantic orchestration, and output validation to ensure accuracy and repeatability in data extraction. It employs techniques such as temperature tuning, multi-model ensembles, and human-in-the-loop reviews to validate and optimize outputs, while confidence scoring and threshold settings allow users to tailor the system to their risk tolerance. Box Extract continuously learns and improves through user feedback, refining its ground truth datasets and optimizing prompt formulations to enhance performance. This approach not only automates more of the data extraction process but also builds trust by ensuring that AI systems deliver reliable and precise results, aligning with the broader vision of creating trustworthy enterprise AI systems.

Extracting metadata from documents stored in Box is streamlined through a template-driven Agent Skill that utilizes Box AI and Model Context Protocol (MCP) tools to automate the process. The Agent Skill, defined in a reusable SKILL.md file, allows for efficient extraction and writing of metadata by using a simple command to interact with Box's system of record, ensuring that workflows are reusable and scalable across various document types and metadata templates. This method avoids the need for custom API clients or glue code, focusing instead on declarative workflows that maintain Box as the central repository, ensuring metadata fields are populated without overwriting existing data. Once installed, this skill can be executed within a Cursor chat environment, allowing for a dry run to validate logic before applying metadata updates, and provides a structured approach to metadata extraction that enhances search, automation, and downstream processes within Box.

Box Extract is a newly launched AI-powered data extraction tool designed to enhance the processing of unstructured content within enterprises by combining advanced OCR capabilities with state-of-the-art AI models. Unlike traditional OCR and legacy IDP tools that struggle with context and require extensive maintenance, Box Extract offers a scalable solution by understanding language, context, and intent, enabling the automatic extraction of information from various document types and saving it as metadata on Box. The tool utilizes leading AI models such as Google Gemini 3 and OpenAI’s GPT 5.2I, along with techniques like extraction-specific retrieval-augmented generation (RAG) and AI graders to improve data quality iteratively. It provides different extraction agents to balance accuracy and budget needs while supporting enterprise-scale operations through features like Custom Extract Agents and automated workflow integration. Box Extract aims to automate business processes, enhance content discovery, and enable smarter decision-making across industries by delivering high-quality, structured data, significantly improving operational efficiency without increasing headcount.

Box, Inc. has announced the general availability of Box Extract, an AI-powered tool designed to automate workflows and extract valuable insights from unstructured content, like contracts and policy documents, transforming it into structured metadata. Utilizing advanced generative AI models from companies such as Google, Anthropic, and OpenAI, Box Extract aims to enhance decision-making and efficiency across various industries by securely converting unstructured information into actionable data. The tool is particularly beneficial for sectors like financial services, government, media, and insurance, where it can streamline processes such as loan origination and contract management by automatically extracting and organizing critical details. Box Extract's capabilities include creating custom Extract Agents tailored to specific business needs and integrating metadata into third-party applications, facilitating faster search and improved content management. The service is available to Box customers under the Enterprise Advanced plan, providing options for standard and enhanced data capture, depending on document complexity.

Box Extract is a comprehensive AI-based document extraction system designed to handle complex real-world documents by transforming them into structured, queryable data within a secure, enterprise-grade framework. Unlike simple LLM prompting solutions, Box Extract offers a full-stack architecture that includes multiple entry points such as manual user triggers, automatic activation upon file upload, and integration into larger business logic through APIs. The system prioritizes security and compliance by ensuring permissions are verified before processing and utilizes advanced OCR to prepare documents in an AI-friendly format. The Box Extract Agent plays a central role in categorizing, refining, and validating extracted data, with human review steps in place for low-confidence or sensitive extractions. Once processed, the structured data is stored as metadata, enabling enhanced content management, dashboard creation, workflow orchestration, and integration with external systems like Salesforce and Databricks. The platform's infrastructure ensures seamless API consumption, eliminating common errors and adapting to various extraction needs, all while maintaining strict security and compliance standards.

AI content creation offers numerous advantages for businesses, including enhanced data security, increased productivity, and improved scalability by using machine learning models to generate documents such as reports, summaries, and contracts. By integrating AI tools like Box AI into their systems, companies can streamline content creation within a secure environment, ensuring compliance with brand safety and legal standards while preventing data leaks into public training sets. These AI tools excel in understanding context, tone, and intent, producing outputs that are not only grammatically accurate but also relevant to the brand strategy. AI-driven content creation can significantly reduce review cycle times and enable the transformation of unstructured data into actionable insights. Platforms like Box AI provide features such as document generation, data extraction, and content repurposing within a secure framework, leveraging enterprise-grade security and compliance policies. The effectiveness of AI in content creation depends on providing specific context and instructions, allowing businesses to meet large content demands efficiently without increasing headcount.

Enterprises spend significant resources on extracting data from unstructured documents such as contracts and invoices, often leading to high costs and errors when done manually. Generative AI offers a solution by understanding text as humans do, yet this alone is insufficient for effective data extraction. Box has developed innovations that use agentic AI within a content management platform to enhance data extraction accuracy, by employing techniques like named entity recognition and model-based chunk re-ranking. This approach allows for a more focused extraction process, reducing errors and API costs while maintaining compliance and security. Agentic AI systems are capable of self-correction and iterative reasoning, improving accuracy by up to ten percentage points over traditional methods. By integrating data storage and governance within the same platform, Box ensures that the extracted data remains connected to its source, enabling enterprises to efficiently process diverse document types without compromising security. This method not only simplifies compliance but also allows human reviewers to focus on more complex cases, making it economically and operationally viable for large-scale enterprise applications.

Box's Intelligent Prospecting Agent is a new AI tool designed to enhance sales messaging, characterized as a "product marketer in every seller’s back pocket." To develop this agent, Box follows a structured four-phase approach: Ideation, Piloting, Rollout, and Scaled Adoption. The current phase, Piloting, involves testing the tool with 25 users to solve real business problems, gathering feedback, and iterating on its functionality. The process emphasizes a minimum viable product (MVP) approach, where feedback from test users is crucial for refining the tool without heavy investment. Key roles in the pilot include functional leaders, an AI manager, and build teams, with test users providing critical feedback. The success of the agent is measured against metrics such as efficiency gains, automation rates, and net-new work enabled. Although not all pilots succeed, they offer valuable insights into AI development. The ongoing pilot phase for the Prospecting Agent has reduced message preparation time and allowed for personalized industry-specific outreach, signaling promising potential as it moves toward wider adoption.

The text explores the impact of artificial intelligence (AI) on the future of work, emphasizing that AI will automate workflows and change job structures rather than replace jobs outright. It highlights how AI will transform roles across industries, necessitating a redesign of work that emphasizes human judgment, creativity, and relationships over repeatable tasks. Gartner research predicts that by 2030, AI will be integral to all work, with a significant portion of tasks being either augmented or fully automated by AI. The article stresses the importance of preparing the workforce for this shift by equipping them with AI literacy and aligning human readiness with technological advancements. It also addresses the expertise paradox, where organizations with strong human domain expertise benefit more from AI. The author, reflecting on leadership in the AI era, advocates for clear communication about AI changes, visible learning, and maintaining human-centric leadership to foster a supportive work environment amidst rapid technological advancements.

AI document extraction is a transformative technology that utilizes artificial intelligence to automate the processing, classification, and extraction of data from diverse document types, such as PDFs, scans, and emails, thereby converting unstructured content into structured, actionable information. This process leverages machine learning, natural language processing (NLP), and optical character recognition (OCR) to effectively handle complex forms, multilingual documents, and varied layouts, reducing human error and enhancing data quality. By integrating AI document extraction into business workflows, organizations can accelerate decision-making, lower operational costs, and improve scalability without increasing headcount. The technology supports enterprise-grade security and governance, facilitating seamless integration with existing business systems like ERPs and CRMs. Solutions like the Box Intelligent Content Management platform offer developer-friendly features, including robust APIs and SDKs, which enable rapid deployment and customization, ultimately transforming data into searchable and governable assets that drive business innovation and efficiency.

Artificial intelligence (AI) is revolutionizing enterprise operations by moving beyond basic automation to achieve complex outcomes, such as automated data extraction, instant document summarization, and accelerated compliance reviews. The guide outlines various AI use cases tailored for specific industries and functions, including generative AI for content creation, AI-assisted process automation, and intent-driven search capabilities. AI tools offer significant efficiency and accuracy improvements by automating time-consuming processes like data extraction and compliance checks, while also enhancing decision-making with real-time insights and scenario analyses. Organizations leverage AI for personalized marketing, cybersecurity, supply chain optimization, and HR management, benefiting from AI's ability to process large datasets, identify patterns, and provide quick, actionable insights. The integration of AI with platforms like Box enables businesses to manage content securely and efficiently through instant data extraction, intelligent workflows, and no-code applications, ensuring seamless collaboration and governance in complex enterprise environments.

AI teamwork integrates artificial intelligence into business processes to enhance collaboration, efficiency, and strategic operations by combining human insight with AI capabilities. This integration is not meant to replace humans but to act as a partner in high-performing teams, leading to increased efficiency, reduced administrative costs, faster time-to-market, and improved compliance. Effective AI teamwork involves centralizing data, adopting a multi-model strategy, automating processes with AI agents, implementing strong governance, and training teams to collaborate with AI. The Box Intelligent Content Management platform exemplifies this by offering AI-powered solutions for various industries, enabling streamlined workflows, enhanced market intelligence, and secure content management. Recent developments focus on embedding AI within existing platforms, offering flexible AI model choices, real-time governance, and no-code tools for custom app creation, all aimed at making AI accessible and tailored to specific enterprise needs.