Recent News in Analytics and AI: November 2025 Edition

This month's edition spotlights Power BI's recognition as a leader in semantic technologies, but that's just the beginning. Microsoft Fabric's Data Agents get a major upgrade with secure, on-demand AI connectivity thanks to an MCP server endpoint, while Copilot Business brings generative AI to small and medium-sized businesses for the first time. The Microsoft Foundry Agent Service quietly debuts memory capabilities, and DuckDB raises the bar for data protection in open-source analytics.

Read on and get up to speed.

Power BI

• Microsoft has been recognised as a leader and outperformer in GigaOm's 2025 Radar for Semantic Layers and Metric Stores. This achievement highlights Power BI's strong capabilities in delivering semantic consistency and trusted metrics across organisations. The report praises Microsoft's integration of semantic layers with AI-driven insights, enabling businesses to unify data governance and accelerate decision-making. Power BI's semantic layer approach ensures accuracy and scalability, supporting enterprise-grade analytics. This recognition reinforces Microsoft's commitment to innovation in data modelling and semantic technologies, positioning Power BI as a top choice for organisations seeking reliable, AI-ready analytics solutions. Learn more.

• Power BI semantic models are emerging as key enablers for AI-driven analytics. By providing a structured, governed layer of business logic, these models allow AI tools to interpret data accurately and deliver meaningful insights. Microsoft emphasises that semantic models bridge the gap between raw data and AI consumption, ensuring consistency and trust in outputs. They also simplify integration with natural language queries and advanced AI features, making analytics more intuitive for users. This approach accelerates decision-making and enhances productivity, positioning semantic models as a cornerstone for organisations adopting AI-powered business intelligence. Learn more.

• Power BI introduces new image visual enhancements to improve report design and user experience. These updates allow creators to customise image placement, scaling, and alignment with greater precision, ensuring visuals complement data storytelling. Features include better support for responsive layouts and accessibility, making reports more engaging and inclusive. Microsoft's goal is to give report authors more creative control while maintaining performance and usability. These enhancements reflect Power BI's commitment to empowering users with tools that combine aesthetics and functionality, helping organisations deliver visually compelling, data-rich reports. Learn more.

• PBIR is set to become the default report format in Power BI, replacing PBIX to support cloud-first and collaborative workflows. Microsoft highlights PBIR's advantages, such as improved versioning, integration with Git-based repositories, and better alignment with enterprise governance standards. The new format enhances interoperability and simplifies report lifecycle management, making it easier for teams to collaborate on analytics projects. Users are encouraged to familiarise themselves with PBIR and plan for migration to ensure a smooth transition. This update underscores Microsoft's commitment to modernising Power BI for scalable, secure, and agile reporting. It‘s important to note that this change is to the internal report format but the PBIX file format will remain the primary format for Power BI files. Learn more.

Microsoft Fabric

• Ignite 2025 introduced major upgrades to Fabric Data Agents, including a Managed MCP Server endpoint that lets AI systems, such as those in VS Code, connect securely to enterprise data for on-demand insights. Microsoft also announced deep integration with Microsoft 365 Copilot, enabling users to access authorised data agents directly from familiar environments like Teams or Office apps. Sharing is streamlined, allowing users to securely share their agents with colleagues for collaborative analytics. Combined with enhanced reasoning and AI interoperability, these updates position Data Agents as a bridge between enterprise data and productivity, driving faster, smarter decision-making. Learn more.

• Microsoft Fabric IQ redefines analytics by embedding intelligence and ontology-driven reasoning into the platform. Ontologies provide structured business context, enabling AI to interpret relationships and deliver accurate, meaningful insights. Fabric IQ combines this with natural language querying, predictive modelling, and automated recommendations, making analytics more intuitive. Users can interact conversationally with data while maintaining governance and compliance. By integrating ontology with AI, Fabric IQ moves beyond raw data to deliver contextual intelligence, helping organisations accelerate decision-making and innovation. This evolution positions Fabric as a true intelligence platform rather than just a data repository. Learn more.

• Microsoft announced the preview of dbt Jobs in Fabric, enabling teams to build and deploy SQL-based models more efficiently. This integration allows users to orchestrate dbt workflows directly within Fabric, reducing complexity and improving governance. By leveraging dbt's transformation capabilities, organisations can ensure data models are accurate, version-controlled, and aligned with best practices. The feature accelerates development cycles and supports collaborative analytics, making it easier to maintain trustworthy data pipelines. This preview demonstrates Microsoft's commitment to combining modern data engineering tools with Fabric's unified analytics environment. Learn more.

• OneLake now offers fine-grained ReadWrite permissions, allowing organisations to separate data write access from broader workspace control. Previously, users needing write access also had extensive permissions to manage artefacts. The new model supports all write operations (including upload, delete, rename, and edit) through Spark notebooks, OneLake File Explorer, and APIs. Shortcut CRUD operations are also enabled, making it easier to share data without creating copies. Write access via Lakehouse UX for viewers is not yet available but will be added soon. This update improves security, flexibility, and collaboration by introducing more precise access controls for modern lakehouse environments. Learn more.

Microsoft 365 Copilot and Copilot Studio

• Microsoft has launched Copilot Business, a tailored version of Microsoft 365 Copilot designed for small and medium-sized businesses (SMBs). This offering brings AI-powered productivity tools to SMBs at an accessible price point, enabling features like natural language document creation, meeting summaries, and workflow automation. Copilot Business integrates seamlessly with Microsoft 365 apps such as Word, Excel, and Teams, helping smaller organisations boost efficiency without requiring advanced technical expertise. With built-in security and compliance, it ensures data protection while democratising AI capabilities. This launch reflects Microsoft's commitment to making generative AI practical and affordable for businesses of all sizes. Learn more.

• OpenAI's Sora 2 is now available in Microsoft 365 Copilot, introducing AI-driven video generation to everyday productivity tools. Users can create dynamic, high-quality videos for presentations, campaigns, or learning materials without leaving apps like PowerPoint or Teams. This integration enhances Copilot's creative capabilities, complementing its text and image generation features. Microsoft ensures that Sora 2 operates within secure, compliant frameworks, making it suitable for enterprise use. By adding video generation, Microsoft positions Copilot as a comprehensive AI assistant for content creation, helping organisations communicate more effectively and creatively. Learn more.

• Microsoft Copilot Studio now supports GPT-5.1, the latest generation of OpenAI's language model, delivering improved reasoning, accuracy, and conversational capabilities. This upgrade enhances custom Copilot development, enabling creators to build more intelligent, context-aware assistants tailored to business needs. GPT-5.1 offers better performance in complex tasks, multi-turn dialogues, and natural language understanding, making Copilot Studio a powerful platform for enterprise AI solutions. Microsoft highlights that the integration maintains strong security and compliance, ensuring responsible AI deployment. With GPT-5.1, organisations can create advanced copilots that drive productivity and innovation across diverse workflows. Learn more.

• Microsoft Copilot Studio now supports Claude Opus 4.5, Anthropic's latest AI model, offering enhanced reasoning and safety features. Claude Opus 4.5 excels in handling complex tasks and delivering reliable, context-aware responses, making it ideal for enterprise-grade copilots. The integration gives developers more choice, enabling them to combine Claude's strengths with Microsoft's secure, compliant environment. This update reflects Microsoft's commitment to providing diverse, advanced AI models for building tailored solutions. With Claude Opus 4.5, organisations can create copilots that prioritise accuracy, transparency, and responsible AI practices. Learn more.

Azure Analytics and AI

• The introduction of memory in the Microsoft Foundry Agent Service marks a significant step towards more intelligent AI agents. Memory allows agents to retain context from previous interactions, enabling them to provide personalised responses and handle complex workflows with greater accuracy. Developers can configure memory settings to align with organisational needs, ensuring flexibility and compliance. This capability reduces repetitive queries and enhances multi-step reasoning, improving overall efficiency. By embedding memory into its agent framework, Microsoft Foundry empowers businesses to build AI solutions that deliver consistent, context rich experiences across applications and services. Learn more.

• Multi-agent workflows are now available in Microsoft Foundry Agent Service as a public preview, enabling coordinated execution of complex tasks by multiple agents. This capability allows developers to create workflows where agents collaborate, share context, and divide responsibilities for greater efficiency. It is particularly useful for scenarios involving data analysis, automation, and operational decision-making. By supporting multi-agent orchestration, Foundry enhances its ability to deliver scalable and intelligent AI solutions. Organisations can now design systems that combine specialised expertise across agents, improving accuracy and accelerating outcomes in enterprise environments. Learn more.

• Databricks announced expanded support for OneLake through Unity Catalog, enabling unified governance across data platforms. This integration allows organisations to manage permissions, lineage, and compliance consistently for data stored in OneLake and Databricks environments. Unity Catalog provides a centralised approach to security and metadata management, reducing complexity and improving collaboration. By connecting OneLake with Unity Catalog, businesses can streamline data sharing and maintain trust in analytics workflows. This update strengthens interoperability between Microsoft Fabric and Databricks, supporting enterprise strategies for secure and scalable data management. Learn more.

• Microsoft Ignite 2025 showcased major Azure innovations focused on AI, security, and cloud efficiency. Highlights include expanded AI services, improved integration with Microsoft Fabric, and new tools for building intelligent applications. Azure introduced features that simplify deployment of AI models, enhance governance, and optimise performance across hybrid and multi-cloud environments. Security updates strengthen identity management and compliance, while new developer tools accelerate application development. These announcements reflect Microsoft's vision of an intelligent cloud that combines scalability, security, and advanced AI capabilities to empower organisations globally. Learn more.

Open-Source Analytics and AI

• DuckDB now offers built-in encryption, providing robust protection for data stored in its analytical engine. Users can encrypt database files and backups using strong cryptographic standards, ensuring confidentiality and compliance. The feature includes options for external key management, allowing integration with enterprise security frameworks. Secure connections are also supported, reducing risks during data transfer. This enhancement makes DuckDB more appealing for sensitive workloads, combining security with its hallmark simplicity and performance. With encryption, DuckDB delivers a secure foundation for analytics without introducing unnecessary complexity. Learn more.

• Polars is at the heart of Citizens' real-time analytics strategy, replacing slower legacy systems with a high-performance solution. Built on Rust and optimised for columnar data, Polars delivers lightning-fast computations for large-scale datasets. Citizens uses Polars to power dashboards and operational insights, enabling teams to respond quickly to changing conditions. The platform integrates easily with existing pipelines, reducing complexity and accelerating deployment. This case study showcases Polars' ability to combine speed, scalability, and simplicity, making it a compelling choice for organisations seeking real-time analytics capabilities. Learn more.

• Daft AI argues that agentic systems, often marketed as autonomous AI, are essentially query engines for unstructured data. These systems excel at retrieving and synthesising information from diverse sources but rely heavily on structured prompts and orchestration. While they appear autonomous, most workflows involve executing queries across APIs and databases rather than true reasoning. The article calls for clarity in defining agentic capabilities, positioning them as advanced retrieval tools rather than independent agents. Understanding this distinction helps organisations set realistic expectations and design workflows that leverage strengths without overestimating autonomy. Learn more.

• Meta has introduced SAM 3D, an evolution of its Segment Anything Model designed for three-dimensional understanding. SAM 3D extends segmentation capabilities from 2D images to 3D environments, enabling applications in robotics, AR, and spatial computing. The model uses advanced vision algorithms to interpret depth and geometry, delivering accurate object segmentation in complex scenes. Meta highlights SAM 3D's potential for tasks like autonomous navigation and immersive experiences, where spatial awareness is critical. This innovation marks a significant step in bridging computer vision and 3D perception, opening new possibilities for AI-driven interaction with physical spaces. Learn more.

Stay in the Know

Get notified when we post something new by following us on X and LinkedIn.