The Biggest Announcements from Microsoft Ignite 2025
Hi, everyone! I've spent the last day sorting everything from Microsoft Ignite and have together this overview of the headline announcements, along with a number of other notable ones.
As always, the Microsoft Ignite 2025 Book of News is a great entry point to everything Microsoft, where I put more focus on Azure and Fabric.
Foundry Control Plane: Unified Operations and Governance for AI Agents
The Foundry Control Plane is a new centralized platform designed to help organizations build, operate, and govern their entire fleet of AI agents from a single location. It addresses the "sprawl" of agents by providing a unified view regardless of where the agent was built (e.g., Microsoft Foundry, LangChain, or external frameworks).
- Governance & Security: Integrates with Microsoft Entra to assign unique IDs to agents, ensuring they are tracked and governed like employees. It also leverages Microsoft Defender and Purview to enforce security policies, block threats, and prevent data leakage.
- Observability: Provides developers with deep visibility into agent behavior, including performance metrics, cost tracking, and tracing of every step an agent takes.
- Guardrails: Allows teams to set strict controls on agent inputs, outputs, and tool interactions to ensure they remain safe, accurate, and compliant with business rules.

Learn more: Foundry Control Plane: Where Developers Build, Operate, and Govern Every Agent
Foundry IQ: A Unified Knowledge Layer for Agents
Foundry IQ simplifies how agents access and use enterprise data, replacing complex, fragmented "Retrieval-Augmented Generation" (RAG) pipelines with a single, managed knowledge layer. It ensures agents have access to trusted, up-to-date information while respecting security permissions.
- Agentic Retrieval: Uses an advanced AI engine that plans, searches, and synthesizes information from across the organization to answer complex queries, rather than just performing simple keyword lookups.
- Unified Data Access: Connects agents to a wide range of data sources—including Microsoft 365 SharePoint, Fabric OneLake, Azure AI Search, and the web—through reusable "knowledge bases."
- Security First: Honors existing document-level permissions (e.g., Entra ID), ensuring agents never reveal information a user isn't authorized to see.
Learn more:
- Foundry IQ: Unlock knowledge for Agents
- Foundry IQ: boost response relevance by 36% with agentic retrieval
Fabric IQ: Transforming Data into Business Intelligence
Fabric IQ elevates Microsoft Fabric from a data platform to an "intelligence platform," enabling organizations to model their business in a way that both humans and AI agents can understand and act upon.
- Business Ontology: Allows business experts (not just engineers) to create a shared digital model of their business concepts (e.g., defining what a "customer" or "order" is) and how they relate to one another.
- Operations Agents: Introduces autonomous agents that monitor live data streams to detect risks or opportunities and take action in real-time without human intervention.
- Cross-Platform Intelligence: Integrates with Foundry IQ and Work IQ to create a shared intelligence layer, ensuring that data insights in Fabric are available to agents working in Microsoft 365 or custom applications.
Learn more: From Data Platform to Intelligence Platform: Introducing Microsoft Fabric IQ
Azure HorizonDB: High-Performance Postgres for the AI Era

Azure HorizonDB is a new, fully managed PostgreSQL-compatible database service built for high-scale enterprise and AI workloads. It is designed to handle the intense demands of modern applications while simplifying migration from legacy systems.
- Massive Scalability: Features a cloud-native architecture that separates compute from storage, allowing for independent scaling up to 3,072 vCores and 128TB of storage.
- AI Optimization: Includes built-in support for vector search (critical for AI applications) and seamless integration with Microsoft Foundry models, making it easier to build intelligent apps directly on the database.
- Migration Tools: Offers GitHub Copilot-powered capabilities to automate the conversion of complex code, simplifying the migration of legacy databases (like Oracle) to the cloud.
Learn more: Announcing Azure HorizonDB
Azure Copilot: Agentic Cloud Operations

Azure Copilot is evolving from a chat assistant into an "agentic" interface that actively manages cloud operations on your behalf. It orchestrates specialized agents to handle complex tasks across the entire cloud lifecycle, moving from reactive management to proactive automation.
- Specialized Agents: Deploys dedicated agents for specific tasks including Migration, Deployment, Observability, Optimization, Resiliency, and Troubleshooting.
- Autonomous Action: Can perform multi-step reasoning to plan and execute complex workflows—such as diagnosing a VM issue and applying a fix—while always asking for human approval before making changes.
- Unified Management: Accessible directly within the Azure Portal, CLI, and console, it provides a personalized experience that understands your specific resources and permissions.
Learn more:
- Ushering in the Era of Agentic Cloud Operations with Azure Copilot
- Migration Agent - Unlocking transformation
Microsoft Agent 365: The Control Plane for Workplace Agents

Microsoft Agent 365 acts as the central command center for managing AI agents within the Microsoft 365 ecosystem. It provides IT leaders with the tools needed to deploy and oversee agents securely alongside their human workforce.
- Central Registry: Offers a single source of truth (via Microsoft Entra) to inventory every agent in the organization, preventing "shadow AI."
- Compliance & Security: Enforces strict access controls and policies, utilizing Microsoft Defender to block attacks and Microsoft Purview to prevent sensitive data exposure.
- ROI & Visualization: Includes dashboards to visualize agent connections and measure their performance and business impact, helping organizations track the return on investment for their AI tools.
Learn more: Announcing New Microsoft Purview Capabilities to Protect GenAI Agents

Other Notable Announcements
Besides the headline announcements above, there were a bunch more that are worth your attention.
SQL Server 2025: The AI-Ready Enterprise Database

SQL Server 2025 is now generally available, marking a significant evolution of the platform by integrating native AI capabilities directly into the database engine while enhancing its security and hybrid cloud performance.
- Built-in AI: Features native support for vector embeddings and text chunking, allowing organizations to run secure AI workloads (like RAG and semantic search) directly within SQL Server without moving data.
- Developer Productivity: Introduces native JSON support, REST APIs, and regular expressions (RegEx) to streamline app development, along with GitHub Copilot integration for SQL Server Management Studio.
- Hybrid Agility: Enhances Azure Arc integration for unified governance across on-premises and cloud environments, and database mirroring in Microsoft Fabric for real-time analytics without ETL.
- Performance & Scale: Delivers optimized locking to reduce memory usage and boost concurrency, alongside increased resource limits for Standard Edition (up to 32 cores and 256 GB memory).
Learn more: SQL Server 2025 is Now Generally Available
Azure DocumentDB: Fully Managed MongoDB-Compatible Database
Azure DocumentDB (formerly Azure Cosmos DB for MongoDB vCore) is now generally available as a fully managed, MongoDB-compatible database service. It's built on open-source foundations to ensure high portability and no vendor lock-in, making it an ideal choice for modernizing existing MongoDB workloads.
- Open Source & Multi-Cloud: Based on the Linux Foundation-governed open-source DocumentDB project, allowing the same engine to run locally, on-premises, or across multiple clouds (Azure, AWS, GCP) via Kubernetes.
- Predictable Pricing: Uses a simple, vCore-based pricing model with independent scaling for compute and storage, offering clear cost control and transparency (no separate costs for backups).
- Enterprise Ready: A first-party Azure service with native Microsoft Entra ID authentication, customer-managed keys (CMK), and a full-stack 99.995% SLA that covers compute and networking.
- Seamless Migration: Supported by built-in Azure migration tools and standard MongoDB tools, allowing developers to move workloads easily while retaining their existing skills and tooling.

Anthropic's Claude models in Microsoft Foundry
The Claude models are currently the state-of-the-art frontier LLMs for coding, complex development, and long-running agents. It's great to see these as first class options in Microsoft Foundry next to the OpenAI GPT models.

Security Copilot now included with Microsoft 365 E5
Security Copilot has been around for a couple of years as a separate service and is now included as a capability within the M365 E5 subscription.

This change will allow more customers to take advantage of the advanced features of Security Copilot with their current investment in the E5 suite.

Foundry Tools: Connecting Agents to the Enterprise World Foundry
Foundry Tools is a new unified catalog within Microsoft Foundry that empowers developers to equip AI agents with the capabilities they need to take meaningful action. By standardizing how agents discover and use tools via the Model Context Protocol (MCP), Microsoft is changing agents from simple conversationalists into powerful automation systems.
- Massive Tool Ecosystem: Instantly provides agents with access to over 1,400 Azure Logic Apps connectors and the full portfolio of Azure AI Services (like Speech and Document Intelligence), all treated as "tools" that agents can invoke.
- Seamless Integration: Enables agents to interact securely with core business systems—including Outlook, Teams, SharePoint, and Dataverse—allowing them to automate tasks like scheduling meetings, drafting emails, and updating CRM records.
- Secure Governance: Features a private catalog (via Azure API Center) where organizations can publish and manage their own internal tools, ensuring that all agent actions are traceable, compliant, and secured with Microsoft Entra ID.
Learn more: Build smarter, connected agentic apps with Foundry Tools
Microsoft Agent Factory: Accelerating the Move from Experimentation to Execution
Microsoft Agent Factory is a new program designed to help organizations overcome the barriers of procurement complexity and skills gaps, allowing them to build and deploy AI agents faster and more predictably.
- Unified Metered Plan: Introduces the "Microsoft Agent Pre-Purchase Plan (P3)," which allows customers to purchase Agent Commit Units (ACUs) upfront. These units can be used flexibly across both Copilot Studio and Microsoft Foundry services, simplifying procurement and eliminating the need to manage multiple contracts.
- Forward Deployed Engineers: Eligible organizations can partner directly with Microsoft's AI experts (Forward Deployed Engineers) to co-innovate and build production-ready solutions in weeks, while retaining full ownership of their data and intellectual property.
- Customized Skilling: Includes tailored, role-based training programs to boost AI fluency across the organization, ensuring that teams are equipped to adopt and scale agentic innovation effectively.
- Deployment Flexibility: Enables makers to build agents infused with Work IQ, Fabric IQ, and Foundry IQ and deploy them anywhere (including Microsoft 365 Copilot) without upfront licensing hurdles.
Learn more: Introducing Microsoft Agent Factory

Fabric Databases: Unified SaaS-Native Experience for Modern Workloads
Fabric Databases are now generally available, merging the power of SQL Database and Cosmos DB into a single, unified, SaaS-native platform within Microsoft Fabric. This release streamlines the management of both transactional and analytical data, allowing organizations to build AI-powered apps faster and with less complexity.
- Unified Platform: Combines SQL Database (relational) and Cosmos DB (NoSQL) in one environment, allowing developers to query live operational data without complex ETL pipelines.
- AI & App Innovation: Features native AI integration, including support for vector search and RAG patterns, making it easier to build intelligent agents and applications that can read and write data in real-time.
- Translytical Capabilities: Enables "translytical" workflows where operational tasks (like updating CRM records) and analytical insights happen in the same space, allowing for immediate action on data insights directly from dashboards.
- Enterprise-Grade & Serverless: Offers a fully autonomous, serverless architecture that scales automatically, backed by unified security and billing to simplify operations at a global scale.

Managed Instance on Azure App Service: Streamlining Migration for Legacy & Complex Apps
Managed Instance on Azure App Service is a new capability, currently in Public Preview, designed to help organizations migrate and modernize legacy applications—particularly .NET Framework apps—without requiring major code rewrites. It bridges the gap between traditional infrastructure and modern PaaS.
- Legacy Compatibility: Overcomes common migration blockers by allowing apps to access file systems, Windows Registry settings, and custom OS-level dependencies (like GAC entries or Windows Services) directly on the instance.
- Simplified Configuration: Features "configuration scripts" that automatically install and configure necessary dependencies from a zip file, along with registry adapters that securely manage settings via Azure Key Vault.
- Enhanced Access: Offers unprecedented access for a PaaS service, including the ability to RDP into instances via Azure Bastion for troubleshooting with familiar tools like IIS Manager and Event Viewer.
- Cost & Speed: Accelerates cloud adoption by reducing the need for expensive re-platforming, allowing businesses to lift-and-shift hard-to-modernize apps and realize ROI faster while still benefiting from Azure’s scaling and security.
Learn more:
- Announcing Public Preview of Managed Instance on Azure App Service
- Managed Instance on App Service overview (preview)
A Collection of Other Microsoft Fabric Announcements

- SAP Business Data Cloud Connect for Microsoft Fabric: Zero-copy, bi-directional sharing between SAP Business Data Cloud and Microsoft Fabric. This will allow businesses to combine semantically rich SAP data with non-SAP data in Microsoft OneLake for advanced analytics with Power BI and easier integration with AI apps and agents.
- Enhanced Interoperability Between Microsoft Fabric and Snowflake: Seamlessly access all Snowflake data within Fabric using any workload — analytics, AI, or visualization — without extra configuration.
- Native reading of OneLake through Azure Databricks Unity Catalog: Seamlessly access data stored in OneLake from Azure Databricks without duplication or complex configuration.
- Native Integration of dbt Fusion with Fabric Data Factory: Allows data engineers to easily use dbt-based data transformations within Fabric for a truly code-based experience that makes it easier to test, version control, and reuse.
- OneLake Shortcuts to SharePoint and OneDrive: Allows you to easily bring unstructured data into OneLake without copying or configuring ETL pipelines to make it easier to train AI agents and enrich structured data.
- New Capabilities in Fabric Data Agents: Fabric data agents now support mirrored databases, SQL database in Fabric, unstructured data, and Ontology as data and knowledge sources. Data agents also now have much better integration with other AI tools via a hosted MCP server and interoperability with Microsoft 365 Copilot.



