Orchestrate multimodal AI insights within your healthcare data estate (Public Preview)
In today’s healthcare landscape, there is an increasing emphasis on leveraging artificial intelligence (AI) to extract meaningful insights from diverse datasets to improve patient care and drive clinical research. However, incorporating AI into your healthcare data estate often brings significant costs and challenges, especially when dealing with siloed and unstructured data. Healthcare organizations produce and consume data that is not only vast but also varied in format—ranging from structured EHR entries to unstructured clinical notes and imaging data. Traditional methods require manual effort to prepare and harmonize this data for AI, specify the AI output format, set up API calls, store the AI outputs, integrate the AI outputs, and analyze the AI outputs for each AI model or service you decide to use. Orchestrate multimodal AI insights is designed to streamline and scale healthcare AI within your data estate by building off of the data transformations in healthcare data solutions in Microsoft Fabric. This capability provides a framework to generate AI insights by connecting your multimodal healthcare data to an ecosystem of AI services and models and integrating structured AI-generated insights back into your data estate. When you combine these AI-generated insights with the existing healthcare data in your data estate, you can power advanced analytics scenarios for your organization and patient population. Key features: Metadata store lakehouse acts as a central repository for the metadata for AI orchestration to effectively capture and manage enrichment definitions, view definitions, and contextual information for traceability purposes. Execution notebooks define the enrichment view and enrichment definition based on the model configuration and input mappings. They also specify the model processor and transformer. The model processor calls the model API, and the transformer produces the standardized output while saving the output in the bronze lakehouse in the Ingest folder. Transformation pipeline to ingest AI-generated insights through the healthcare data solutions medallion lakehouse layers and persist the insights in an enrichment store within the silver layer. Conceptual architecture: The data transformations in healthcare data solutions in Microsoft Fabric allow you ingest, store, and analyze multimodal data. With the orchestrate multimodal AI insights capability, this standardized data serves as the input for healthcare AI models. The model results are stored in a standardized format and provide new insights from your data. The diagram below shows the flow of integrating AI generated insights into the data estate, starting as raw data in the bronze lakehouse and being transformed to delta tables in the silver lakehouse. This capability simplifies AI integration across modalities for data-driven research and care, currently supporting: Text Analytics for health in Azure AI Language to extract medical entities such as conditions and medications from unstructured clinical notes. This utilizes the data in the DocumentReference FHIR resource. MedImageInsight healthcare AI model in Azure AI Foundry to generate medical image embeddings from imaging data. This model leverages the data in the ImagingStudy FHIR resource. MedImageParse healthcare AI model in Azure AI Foundry to enable segmentation, detection, and recognition from imaging data across numerous object types and imaging modalities. This model uses the data in the ImagingStudy FHIR resource. By using orchestrate multimodal AI insights to leverage the data in healthcare data solutions for these models and integrate the results into the data estate, you can analyze your existing data alongside AI enrichments. This allows you to explore use cases such as creating image segmentations and combining with your existing imaging metadata and clinical data to enable quick insights and disease progression trends for clinical research at the patient level. Get started today! This capability is now available in public preview, and you can use the in-product sample data to test this feature with any of the three models listed above. For more information and to learn how to deploy the capability, please refer to the product documentation. We will dive deeper into more detailed aspects of the capability, such as the enrichment store and custom AI use cases, in upcoming blogs. Medical device disclaimer: Microsoft products and services (1) are not designed, intended or made available as a medical device, and (2) are not designed or intended to be a substitute for professional medical advice, diagnosis, treatment, or judgment and should not be used to replace or as a substitute for professional medical advice, diagnosis, treatment, or judgment. Customers/partners are responsible for ensuring solutions comply with applicable laws and regulations. FHIR® is the registered trademark of HL7 and is used with permission of HL7.Elevating care management analytics with Copilot for Power BI
Healthcare data solutions care management analytics capability offers a comprehensive template using the medallion Lakehouse architecture to unify and analyze diverse data sets of meaningful insights. This enables enhanced care coordination, improved patient outcomes, and scalable, sustainable insights. As the healthcare industry faces rising costs and growing demand for personalized care, data and AI are becoming critical tools. Copilot for Power BI leads this shift, blending AI-driven insights with advanced visualization to revolutionize care delivery. What is Copilot for Power BI? Copilot is an AI-powered assistant embedded directly into Power BI, Microsoft's interactive data visualization platform. By leveraging natural language processing and machine learning, Copilot helps users interact with their data more intuitively whether by asking questions in plain English, generating complex calculations, or uncovering patterns that might otherwise go unnoticed. Copilot for Power BI is embedded within healthcare data solutions, allowing care management—one of its core capabilities—to harness these AI-driven insights. In the context of care management analytics, this means turning a sea of clinical, claims, and operational data into actionable insights without needing to write a single line of code. This empowers teams across all technical levels to gain value from data. Driving better outcomes through intelligent insights in care management analytics The Care Management Analytics solution, built on the Healthcare data solutions platform, leverages Power BI with Copilot embedded directly within it. Here’s how Copilot for Power BI is revolutionizing care management: Enhancing decision-making with AI Traditionally, deriving insights from healthcare data required technical expertise and hours of analysis. Copilot simplifies this by allowing care managers and clinicians to ask questions like “Analyze which medical conditions have the highest cost and prevalence in low-income regions.” The AI interprets these queries and responds with visualizations, trends, and predictions—empowering faster, data-driven decisions. Proactive care planning By analyzing historical and real-time data, Copilot helps identify at-risk patients before complications arise. This enables care teams to intervene earlier, design more personalized care plans, and ultimately improve outcomes while reducing unnecessary hospitalizations. Operational efficiency From staffing models to resource allocation, Copilot provides visibility into operational metrics that can drive significant efficiency gains. Healthcare leaders can quickly identify bottlenecks, monitor key performance indicators (KPIs) and simulate “what-if” scenarios, enabling more i nformed, data-backed decisions on care delivery models. Reducing costs without compromising quality Cost containment is a constant challenge in healthcare. By highlighting areas of high spend and correlating them with clinical outcomes, Copilot empowers organizations to optimize care pathways and eliminate inefficiencies ensuring patients receive the right care at the right time, without waste. Democratizing data access Perhaps one of the most transformative aspects of Copilot is how it democratizes access to analytics. Non-technical users from care coordinators to nurse managers can interact with dashboards, explore data, and generate insights independently. This cultural shift encourages a more data-literate workforce and fosters collaboration across teams. Real-world impact Consider a healthcare system leveraging Power BI and Copilot to manage chronic disease populations more effectively. By combining claims data, social determinants of health (SDoH) indicators, and patient-reported outcomes, care teams can gain a comprehensive view of patient needs- enabling more coordinated care and proactively identifying care gaps. With these insights, organizations can launch targeted outreach initiatives that reduce avoidable emergency department (ED) visits, improve medication adherence, and ultimately enhance outcomes. The future is here The integration of Copilot for Power BI marks a pivotal moment for healthcare analytics. It bridges the gap between data and action, bringing AI to the frontlines of care. As the industry continues to embrace value-based care models, tools like Copilot will be essential in achieving the triple aim: better care, lower costs, and improved patient experience. Copilot is more than a tool — it is a strategic partner in you care transformation journey. Deployment of care management analytics Showcasing how a Population Health Director uncovers actionable insights through Copilot Note: To fully leverage the capabilities of the solution, please follow the deployment steps provided and use the sample data included with the Healthcare Data Solution. For more information on care management analytics, please review our detailed documentation and get started with transforming your healthcare data landscape today Overview of care management analytics - Microsoft Cloud for Healthcare | Microsoft Learn Deploy and analyze using Care management analytics - Training | Microsoft Learn. Medical device disclaimer: Microsoft products and services (1) are not designed, intended or made available as a medical device, and (2) are not designed or intended to be a substitute for professional medical advice, diagnosis, treatment, or judgment and should not be used to replace or as a substitute for professional medical advice, diagnosis, treatment, or judgment. Customers/partners are responsible for ensuring solutions comply with applicable laws and regulations.
Upgrade performance, availability and security with new features in Azure Database for PostgreSQL
At Microsoft Build 2025 the Postgres on Azure team is announcing an exciting set of improvements and features for Azure Database for PostgreSQL. One area we are always focused on is the enterprise. This week we are delighted to announce improvements across the enterprise pillars of Performance, Availability and Security. In addition, we're improving Integration of Postgres workloads with services like ADF and Fabric. Here's a quick tour of the enterprise enhancements to Azure Database for PostgreSQL being announced this week. Performance and scale SSD v2 with HA support - Public Preview The public preview of zone-redundant high availability (HA) support for the Premium SSD v2 storage tier with Azure Database for PostgreSQL flexible server is now available. You can now enable High Availability with zone redundancy using Azure Premium SSD v2 when deploying flexible server, helping you achieve a Recovery Point Objective (RPO) of zero for mission-critical workloads. Premium SSD v2 offers sub-millisecond latency and outstanding performance at a low cost, making it ideal for IO-intensive, enterprise-grade workloads. With this update, you can significantly boost the price-performance of your PostgreSQL deployments on Azure and improve availability with reduced downtime during HA failover. The key benefits of SSD v2 include: Flexible disk sizing from 1 GiB to 64 TiB, with 1-GiB increment support Independent performance configuration: scale up to 80,000 IOPS and 1,200 MBps throughput without needing to provision larger disks To learn more about how to upgrade and best practices, visit: Premium SSDv2 PostgreSQL 17 Major Version Upgrade – Public Preview PostgreSQL version 17 brings a host of performance improvements, including a more efficient VACUUM process, faster sequential scans via streaming IO, and optimized query execution. Now, with the public preview of in-place major version upgrades to PostgreSQL 17 there is an easier path to v17 for your existing flexible server workloads. With this release, you can upgrade from earlier versions (14, 15, or 16) to PostgreSQL 17 without the need to migrate data or change server endpoints, simplifying the upgrade process and minimizing downtime. Azure’s in-place upgrade capability offers a native, low-disruption upgrade path directly from the Azure Portal or CLI. For upgrade steps and best practices, check out our detailed blog post. Availability Long-Term Backup (LTR) for Azure Database for PostgreSQL flexible server - Generally Available Long-term backups are essential for organizations with regulatory, compliance, and audit-driven requirements, especially in industries like finance and healthcare. Certifications such as HIPAA often mandate data retention periods up to 10 years, far exceeding the default 35-day retention limit provided by point-in-time restore (PITR) capabilities. Long-term backup for Azure Database for PostgreSQL flexible server, powered by Azure Backup is now generally available. With this release, you can now benefit from: Policy-driven, one-click enablement of long-term backups Resilient data retention across Azure Storage tiers Consumption-based pricing with no egress charges Support for restoring backups well beyond community-supported PostgreSQL versions This LTR capability uses a logical backup approach based on pg_dump and pg_restore, offering a flexible, open-source format that enhances portability and ensures your data can be restored across a variety of environments including Azure VMs, on-premises, or even other cloud providers. Learn more about long term retention: Backup and restore - Azure Database for PostgreSQL flexible server Azure Databases for PostgreSQL flexible server Resiliency Solution accelerator When it comes to ensuring business continuity, your database infrastructure is the most critical component. In addition to product documentation, it is important to have access to opinionated solution architecture, industry-proven recommended practices, and deployable infra-as-code that you can learn and customize to ensure an automated production-ready resilient infrastructure for your data. The Azure Database for PostgreSQL Resiliency Solution Accelerator is now available, providing a set of deployable architectures to ensure business continuity, minimize downtime, and protect data integrity during planned and unplanned events. In additional to architecture and recommended practices, a customizable Terraform deployment workflow is provided. Learn more: Azure Database for PostgreSQL Resiliency Solution Accelerator Security Automatic Customer Managed Key (CMK) version updates - Generally Available Azure Database for PostgreSQL flexible server data is fully encrypted, supporting both Service Managed and Customer Managed encryption keys (CMK). Automatic version updates for CMK (also known as “versionless keys”) is now generally available. This change simplifies the key lifecycle management by allowing PostgreSQL to automatically adopt new keys without needing manual updates. Combined with Azure Key Vault's auto-rotation feature this significantly reduces the management overhead of encryption key maintenance. Learn more about automatic CMK version updates. Azure confidential computing SKUs for flexible server - Public Preview Azure confidential computing enables secure sensitive and regulated data, preventing unwanted access of data in-use, by cloud providers, administrators, or external users. With the public preview of Azure confidential SKUs for Azure Database for PostgreSQL flexible server you can now select from a range of Confidential Computing VM sizes to run your PostgreSQL workloads in a hardware-based trusted execution environment (TEE). Azure confidential computing encrypts data in TEE, processing data in a verified environment, enabling you to securely process workloads while meeting compliance and regulatory demands. Learn more about confidential computing with the Azure Database for flexible server. Integration Entra Authentication for Azure Data Factory & Azure Synapse - Generally Available In an era of bring-your-own-device and cloud-enabled apps it is increasingly important for enterprises to maintain central control an identity-based security perimeter. With integrated Entra ID support, Azure Database for PostgreSQL flexible server allows you to bring your database workloads within this perimeter. But how do you securely connect to other services? Entra ID authentication is now supported in the Azure Data Factory and Azure Synapse connectors for Azure Database for PostgreSQL. This feature enables seamless, secure connectivity using Service Principal (key or certificate) and both User-Assigned and System-Assigned Managed Identities, streamlining access to your data pipelines and analytics workloads. Learn more about How to Connect from Azure Data Factory and Synapse Analytics to Azure Database for PostgreSQL. Fabric Data Factory – Upsert Method & Script Activity - Generally Available The Microsoft Fabric has become to go-to data analytics platform with services and tools for every data lifecycle state. To improve customization and fine-grained control over processing of PostgreSQL data, the Upsert Method and custom Script Activity are now generally available in Fabric Data Factory when using Azure Database for PostgreSQL as a source or sink. Upsert Method enables intelligent insert-or-update logic for PostgreSQL, making it easier to handle incremental data loads and change data capture (CDC) scenarios without complex workarounds. Script Activity allows you to embed and execute your own SQL scripts directly within pipelines—ideal for advanced transformations, procedural logic, and fine-grained control over data operations. These capabilities offer enhanced flexibility for building robust, enterprise-grade data workflows, simplifying your ETL processes. Connect to VS Code from the Azure Portal - Public Preview With the exciting announcement of a revamped VS Code PostgreSQL extension preview this week, we're adding a new connection option to the Azure Portal to connect to your flexible server with VS Code, creating a more unified and efficient developer experience. Here's why it matters: One Click Connectivity: No manual connection strings or configuration needed. Faster Onboarding: Go from provisioning a database in Azure to exploring and managing it in VS Code within seconds. Integrated Workflow: Manage infrastructure and development from a single, cohesive environment. Productivity: Connect directly from the Portal to leverage VS Code extension features like query editing, result views, and schema browsing. Where to learn more The Build 2025 announcements this week are just the latest in a compelling set of features delivered by the Azure Database for PostgreSQL team and build on our latest set of monthly feature updates (see: April 2025 Recap: Azure Database for PostgreSQL Flexible Server). Follow the Azure Database for PostgreSQL Blog where you'll see many of the latest updates from Build, including What's New with PostgreSQL @Build, and New Generative AI Features in Azure Database for PostgreSQL.Data security controls in OneLake
Unify and secure your data — no matter where it lives — without sacrificing control using OneLake security, part of Microsoft Fabric. With granular permissions down to the row, column, and table level, you can confidently manage access across engines like Power BI, Spark, and T-SQL, all from one place. Discover, label, and govern your data with clarity using the integrated OneLake catalog that surfaces the right items fast. Aaron Merrill, Microsoft Fabric Principal Program Manager, shows how you can stay in control, from security to discoverability — owning, sharing, and protecting data on your terms. Protect sensitive information at scale. Set precise data access rules — down to individual rows. Check out OneLake security in Microsoft Fabric. No data duplication needed. Hide sensitive columns while still allowing access to relevant data. See it here with OneLake security. Built-in compliance insights. Streamline discovery, governance, and sharing. Get started with the OneLake catalog. QUICK LINKS: 00:00 — OneLake & Microsoft Fabric core concepts 01:28 — Table level security 02:11 — Column level security 03:06 — Power BI report 03:28 — Row level security 04:23 — Data classification options 05:19 — OneLake catalog 06:22 — View and manage data 06:48 — Governance 07:36 — Microsoft Fabric integration 07:59 — Wrap up Link References Check out our blog at https://aka.ms/OneLakeSecurity Sign up for a 60-day free trial at https://fabric.microsoft.com Unfamiliar with Microsoft Mechanics? As Microsoft’s official video series for IT, you can watch and share valuable content and demos of current and upcoming tech from the people who build it at Microsoft. Subscribe to our YouTube: https://www.youtube.com/c/MicrosoftMechanicsSeries Talk with other IT Pros, join us on the Microsoft Tech Community: https://techcommunity.microsoft.com/t5/microsoft-mechanics-blog/bg-p/MicrosoftMechanicsBlog Watch or listen from anywhere, subscribe to our podcast: https://microsoftmechanics.libsyn.com/podcast Keep getting this insider knowledge, join us on social: Follow us on Twitter: https://twitter.com/MSFTMechanics Share knowledge on LinkedIn: https://www.linkedin.com/company/microsoft-mechanics/ Enjoy us on Instagram: https://www.instagram.com/msftmechanics/ Loosen up with us on TikTok: https://www.tiktok.com/@msftmechanics Video Transcript: -As you build AI and analytic workloads, unifying your data from wherever it lives and making it accessible doesn’t have to come at the cost of security. In fact, today we dive deeper into Microsoft’s approach to data unification, accessibility, and security with OneLake, part of Microsoft Fabric, where we’ll focus on OneLake’s security control set and how it compliments data discovery via the new OneLake catalog. -Now, in case you’re new to OneLake and Microsoft Fabric, I’ll start by explaining a few core concepts. OneLake is the logical multi-cloud data lake that is foundational to Microsoft Fabric, Microsoft’s fully managed data analytics and AI platform. OneLake, with its support for open data formats, provides a single and unified place across your entire company for data to be discovered, accessed, and controlled across your data estate. Data can reside anywhere, and you can connect to it using shortcuts or via mirroring. And once in OneLake, you have a single place where data can be centrally classified and labeled as the basis for policy controls. You can then configure granular, role-based permissions that can apply down to the folder level for unstructured data and by table for structured data. -Then all the way down to the column and row levels within each table. This way, security is enforced across all connected data. Meaning that whether you’re accessing the data through Spark, Power BI, T-SQL, or any other engine, it’s protected and you have the controls to allow or limit access to data on your terms. In fact, let me show you a few examples for enforcing OneLake security at all of these levels. I’ll start with an example showing OneLake security at the table level. I want to grant our suppliers team access to a specific table in this lakehouse. I’ll create a OneLake security role to do that. So I’ll just give it a name, SuppliersReaders. Then I’ll choose selected data and find the table that I want to share by expanding the table list, pick suppliers and then confirm. -Now, I just need to assign the right users. I’ll just add Mona in this case, and create the role. Then if I move over to Mona’s experience, I can run queries against the supplier data in the SQL endpoint. But if I try to query any other table, I’m blocked, as you can see here. Now, let me show you another option. This time, I’ll lock access down to the column level. I want to grant our customer relations team access to the data they need, but I don’t want to give them access to PII data. Using OneLake security controls, I can create a role that restricts access to sensitive columns. Like before, I’ll name it. Then I need to select my data. This time, I’ll choose three different tables for customer and order data. But notice this grayed out legacy orders table here that we would like to apply column security to as well. I don’t own the permissions for this table because it’s a shortcut to other data. However, the owner of that data can grant permission to it using the steps I’ll show next. From the role I just created, I’ll expand on my tables. And for the customer’s table, I’ll enable column security. Once I confirm, I can select the columns I want to remove and that we don’t want them to see and save it. -Now, let’s look at the results of this from another engine, Power BI, while building a report. I’ll choose a semantic model for my Power BI report. With the column level security in place, notice the sensitive columns I removed before, contact name and address, are hidden from me. And when I expand the legacy orders table, which was a shortcut, it’s also not showing PII columns. Now, some scenarios require that security controls are applied where records might be interspersed with the same table, so a row level filter is needed. For example, our US-based HR team should only see data for US-based employees. I’ve created another security role with the right data selected, HRUS. -Now, I’ll move to my tables and choose from the options for this employee’s table and I’ll select row security. Row level security in OneLake uses SQL statements to limit what people can see. I’ll do that here with a simple select statement to limit country to USA. Now, from the HR team’s perspective, they can start to query the data using another engine, Spark, to analyze employer retention. But only across US based employees, as you can see from the country column. And as mentioned, this applies to all engines, no matter how you access it, including the Parquet files directly in OneLake. Next, let’s move on to data classification options that can be used to inform policy controls. Here, the good news is the same labels you’ve defined in Microsoft Purview for your organization used in Microsoft 365 for emails, messaging, files, sites, and meetings can be applied to data items in OneLake. -Additionally, Microsoft Purview policy controls can be used to automatically label content in OneLake. And another benefit I can show you from the lineage view is label inheritance. Notice this Lakehouse is labeled Non-Business, as is NorthwindTest, but look at the connected data items on the right of NorthwindTest. They are also non-business. If I move into the test lakehouse and apply a label either automatically or manually to my data, like I’m doing here, then I move back to the lineage view. My downstream data items like this model and the SQL analytics endpoint below it have automatically inherited the upstream label. -So now we’ve explored OneLake security controls, their implementation, and enforcement, let’s look at how this works hand in hand with the OneLake catalog for data discovery and management. First, to know that you’re in the right place, you can use branded domains to organize collections of data. I’ll choose the sales domain. To get the data I want, I can see my items as the ones I own, endorsed items, and my favorites. I can filter by workspace. And on top, I can select the type of data item that I’m looking for. Then if I move over to tags, I can find ones associated with cost centers, dates, or other collection types. -Now, let’s take a look at a data item. This shows me more detail, like the owner and location. I can also see table schemas and more below. I can preview data within the tables directly from here. Then using the lineage tab, it shows me a list of connected and related items. Lastly, the monitor tab lets me track data refresh history. Now, let me show you how as a data owner you can view and manage these data items. From the settings of this lakehouse, I can change its properties and metadata, such as the endorsement or update the sensitivity label. And as the data owner, I can also share it securely internally or even externally with approved recipients. I’ll choose a colleague, [email protected], and share it. -Next, the govern tab in the OneLake catalog gives you even more control as a data owner, as well as recommendations to make data more secure and compliant. You’ll find it on the OneLake catalog main page. This gives me key insights at a glance, like the number and type of items I own. And when I click into view more, I see additional information like my data hierarchy. Below that, item inventory and data refresh status. Sensitivity label coverage gives me an idea of how compliant my data items are. And I can assess data completeness based on whether an item is properly tagged, described, and endorsed across the items I own. Back on the main view, I can see governance actions tailored specifically to my data, like increasing sensitivity label, coverage, and more. -The OneLake catalog is integrated across Microsoft Fabric experiences to help people quickly discover the items they need. And it’s also integrated with your favorite Office apps, including Microsoft Excel, where you can use the get data control to select and access data in OneLake. And right in context, without leaving the app, you can define what you want and pull it directly into your Excel file for analysis. The OneLake catalog is the one place where you can discover the data that you want and manage the data that you own. And combined with OneLake security controls, you can do all of this without increasing your data security risks. -To find out more and get started, check out our blog at aka.ms/OneLakeSecurity. Also, be sure to sign up for a 60 day free trial at fabric.microsoft.com. And keep watching Mechanics for the latest updates across Microsoft, subscribe to our channel, and thanks for watching.The Future of AI: How Lovable.dev and Azure OpenAI Accelerate Apps that Change Lives
Discover how Charles Elwood, a Microsoft AI MVP and TEDx Speaker, leverages Lovable.dev and Azure OpenAI to create impactful AI solutions. From automating expense reports to restoring voices, translating gestures to speech, and visualizing public health data, Charles's innovations are transforming lives and democratizing technology. Follow his journey to learn more about AI for good.Transforming Customer Support with Azure OpenAI, Azure AI Services, and Voice AI Agents
Customer support today is under immense pressure to meet the rising expectations of speed, personalization, and always-on availability. Yet, businesses still struggle with 1. Long wait times and call center 2. queues 3. Disconnected support channels 4. Limited availability of agents outside business hours 5. Repetitive issues consuming valuable human time 6. Frustrated users due to lack of immediate and contextual answers These inefficiencies are costing businesses over $3.7 trillion annually in poor service delivery, while over 70% of agents (based on the research) spend excessive time searching for the right answers instead of resolving problems directly How Voice AI Agents Are Transforming the Support Experience Enter the era of voice-enabled AI agents—powered by Azure OpenAI, Azure AI Services, and ServiceNow—designed to completely transform the way customers engage with support systems. These agents can now: Handle complex user queries in natural language Access enterprise systems (like CRM, ITSM, HR) in real-time Automate repetitive tasks such as password resets, ticket status updates, or return tracking Escalate only when human assistance is truly needed Create connected, seamless, and intelligent support experiences across departments Let’s take a closer look at four architecture patterns that showcase how enterprises can deploy these agents effectively. 🔷 Architecture Pattern 1: Unified Voice Agent with Azure AI + ServiceNow + CRM Integration In this architecture, the customer support journey begins when a user initiates a voice-based conversation through a front-end interface such as a web application, mobile app, or smart device. The captured audio is streamed directly to Azure OpenAI GPT-4o's real-time API, which performs immediate speech-to-text transcription, interprets the intent behind the request, and prepares the initial system response—all in a single seamless stream. Once the user’s intent is understood (e.g., "create a ticket", "check incident status", or "list recent issues"), GPT-4o passes control to Semantic Kernel, which orchestrates the next steps through function calling. Semantic Kernel hosts pre-defined tools (functions) that map to ServiceNow API actions, such as createIncident, getIncidentStatus, listIncidents, or searchKnowledgeBase. These function calls are then securely routed to ServiceNow via REST APIs. ServiceNow executes the appropriate actions—whether it's creating a new support ticket, retrieving the status of an open incident, or searching its Knowledge Base. CRM data is also seamlessly accessed, if needed, to enrich responses with personalized context such as customer history or case metadata. The result from ServiceNow (e.g., an incident ID or KB article summary) is then sent back to Azure GPT-4o, which converts the structured data into a natural spoken response. This final audio output is delivered to the user in real time, completing the end-to-end conversational loop. Additionally, tools like Azure Monitor or Application Insights can be integrated to log telemetry, track usage trends, monitor latency, and analyze user satisfaction over time. This architecture enables organizations to streamline customer support operations, reduce wait times, and deliver natural, intelligent assistance across any channel—voice-first. 🔷 Architecture Pattern 2: Scalable Customer Support with Multi-Agent Voice Architecture This architecture introduces a modular and distributed agent-based design to deliver intelligent, scalable customer support through a voice interface. The process starts with the User Proxy Agent, which acts as the entry point for all user conversations. It captures voice input and forwards the request to the Master Agent, which serves as the brain of the architecture. The Master Agent, empowered with a large language model (LLM) and memory, interprets the intent behind the user's input and dynamically routes the request to the most appropriate domain-specific agent. These include specialized agents such as the Activation Agent, Root Agent, Sales Agent, or Technical Agent, each designed to handle specific workflows or business tasks. The Activation Agent connects to web services and handles provisioning or onboarding scenarios. The Root Agent taps into document search systems (like Azure Cognitive Search) to answer questions grounded in internal documentation. The Sales Agent is equipped with structured logic models (SLMs) and CRM access to retrieve sales-related data from backend databases. The Technical Agent is containerized via Docker and built to manage backend diagnostics, code-level issues, or infrastructure status—often connecting to systems like ServiceNow for real-time ITSM execution. Once the task is executed by the respective agent, results are passed back through the Master Agent and ultimately to the User Proxy Agent, which synthesizes the output into a voice response and delivers it to the user. The presence of shared memory between agents allows for maintaining context across multi-turn conversations, enabling complex, multi-step interactions (e.g., “Create a ticket, check the latest order status, and escalate it if unresolved.”) without breaking continuity. This architecture is ideal for enterprises looking to scale customer support horizontally, adding new agents without disrupting existing workflows. It enables parallelism, specialization, and real-time orchestration, providing faster resolutions while reducing the burden on human agents. Best suited for distributed support operations across IT, HR, sales, and field support—where task-specific intelligence and modular scale are critical. 🔷 Architecture Pattern 3: Customer Support Reinvented with Voice RAG + Azure AI + ServiceNow This architecture brings a cutting-edge twist to Retrieval-Augmented Generation (RAG) by enabling it through a Voice AI agent—creating a truly conversational experience grounded in enterprise knowledge. By combining Azure OpenAI models with the ServiceNow Knowledge Base, this pattern ensures accurate, voice-driven support for employees or customers in real time. The process begins when a user interacts with a voice-enabled interface—via phone, web, or embedded assistant. The Voice AI agent streams the audio to Azure OpenAI GPT-4o, which transcribes the voice input, understands the intent, and then triggers a RAG pipeline. Instead of relying solely on the model’s internal memory, the system performs a real-time query against the ServiceNow Product Knowledge Base, retrieving relevant knowledge articles, troubleshooting guides, or support workflows. These results are embedded directly into the prompt, creating an enriched context that is passed to the language model via Azure AI Foundry. The model then generates a natural, contextually accurate spoken response, which is converted back into audio and voiced to the user—creating a seamless end-to-end Voice RAG experience. This approach ensures that responses are not only conversational but also deeply grounded in trusted enterprise knowledge. Ideal for helpdesk automation, HR support, and IT troubleshooting—where users prefer speaking naturally and need verified, document-backed responses in real time. 🔷 Architecture Pattern 4: Conversational Customer Support with AI Avatars and Azure AI This architecture delivers rich, conversational experiences by integrating AI avatars, Azure AI, and ServiceNow to offer human-like, intelligent customer support across channels. It merges natural speech, facial expression, and enterprise data to create a highly engaging support assistant. The interaction begins when a user speaks with an AI avatar application, whether embedded in a web portal, mobile device, or kiosk. The voice is captured and processed through a speech-to-text pipeline, which feeds the Avatar Module and Live Discussions Engine to manage lip-sync, emotional tone, and turn-taking. Behind the scenes, the avatar is connected to Azure AI services, including Custom Neural Voice (CNV) and Azure OpenAI, which enable the avatar to understand intent and generate responses in natural, conversational language. Most critically, the system integrates directly with the ServiceNow platform. Through secure APIs, the avatar queries ServiceNow to: Retrieve case status updates Provide summaries of incident history Look up Knowledge Base articles Trigger incident creation if needed These ServiceNow results are then passed through the text-to-speech module, with support for multilingual voice synthesis, and rendered by the avatar using expressive animation. Responses are visually delivered as live or pre-rendered avatar videos, creating a truly interactive and personalized experience. This pattern not only answers basic questions but also surfaces dynamic enterprise data—turning the AI avatar into a frontline voice agent capable of real-time, connected support across IT, HR, or customer service domains. Best for branded digital experiences, frontline support stations, or HR/IT helpdesk automation where facial presence, empathy, and backend integration are essential. ✨ Closing Thoughts: The Future of Customer Support Is Here Customer expectations have evolved—and so must the way we deliver support. By combining the power of Azure OpenAI, Azure AI Services, and ServiceNow, we’re not just automating tasks—we’re reinventing how organizations connect with their users. Whether it's: A unified voice agent handling IT tickets and CRM queries, A multi-agent architecture scaling across departments, A voice-enabled RAG system delivering knowledge-grounded answers in real time, or A human-like AI avatar offering face-to-face support— These architectures are driving a new era of intelligent, conversational, and scalable customer service. 👉 Join us at the Microsoft Booth during ServiceNow Knowledge 2025 (starting May 6th) to experience these solutions live, explore the tech behind them, and imagine how they can transform your business. Let’s build the future of support—together.Building Healthcare Research Data Platform using Microsoft Fabric
Co-Authors: Manoj Kumar, Mustafa Al-Durra PhD, Kemal Kepenek, Matt Dearing, Praneeth Sanapathi, Naveen Valluri Overview Research data platforms in healthcare providers, academic medical centers (AMCs), and research institutes support research, clinical decision making, and innovation. They consolidate data from various sources, making it accessible for comprehensive analysis and fostering collaboration among research teams. These platforms automate data collection, processing, and delivery, reducing time and effort needed for data management. This allows researchers to focus on their core activities while ensuring data security and regulatory compliance. The ability to work with multimodal data encourages interdisciplinary and interorganizational collaboration, uniting experts to address complex healthcare challenges. Current challenges Researchers face many common challenges as they work with multimodal healthcare data: Data integration and curation: The process of integrating various data types, such as clinical notes, imaging data, genomic information, and sensor data, presents significant challenges due to differences in formats, standards, and sources. Each AMC employs unique methods for data curation, with some utilizing on-premises solutions and others adopting hybrid cloud systems. No standardized approach currently exists for data curation, necessitating considerable organizational efforts to ensure data consistency and quality. Furthermore, data deidentification is often required to safeguard patient privacy. Data discovery and building cohorts: The lack of a unified multimodal data platform leads to the segregation of data across different modalities. Cohort discovery for each modality is performed separately and often lacks a self-service option, necessitating additional human resources to assist researchers in the data discovery process. This issue is particularly significant because researchers who require Institutional Review Board (IRB) approval cannot access the data beforehand but still need an effective method to identify and explore cohorts. Data delivery: Sensitive patient data, after institutional review board approval, must comply with privacy regulations like the Health Insurance Portability and Accountability Act (HIPAA) and the General Data Protection Regulation (GDPR), requiring secure transfer to prevent breaches. The data, sourced from various systems, needs processing for research readiness. Delivering unified data from modalities like imaging, genomics, and health records is challenging. Typically, research IT teams curate cohort data and deliver it to an SQL database or a file share, accessed by researchers via secure virtual machines. This method often leads to data duplication, creating significant overhead due to numerous ongoing research projects. Cost management: Research projects are funded by government grants and private organizations. Managing these costs is challenging. Research IT departments often implement chargebacks for transparency and accountability in resource use. However, there is a disconnect between funding models and operations. Research teams favor capital expenditure (CapEx) with upfront funding for long-term resources, while cloud platforms operate on operational expenditure (OpEx), incurring ongoing costs based on usage. This shift can lead to concerns about unpredictable costs and budgeting difficulties. Bridging this gap requires careful planning, communication, and hybrid financial strategies to align research needs with cloud-based systems. Compliance with regulations: Healthcare research uses sensitive patient data, requiring strict adherence to HIPAA and GDPR. Transparency in data handling is essential but complex. Researchers must document disclosures thoroughly, detailing who accessed the data and for what purpose. However, tracking and auditing are often fragmented due to inconsistent systems. Variability in disclosure requirements from different agencies adds to compliance challenges. Balancing an auditable trail with privacy and manageable administrative tasks is crucial. Research data platform requirements Ability to curate multi modal data into the research data platform Ability for researchers to identify cohorts (without seeing data) to submit data requests to IRB Automated data delivery after IRB workflow approves the request to access relevant data Tools for researchers as part of the same platform Secure and regulatory-compliant environment for research. An approach to building a research data platform using Microsoft Fabric This article serves as a guide to healthcare organizations, offering a point of view and a prescriptive guidance on building a research data platform using Microsoft Fabric. The solution uses several features from healthcare data solutions in Microsoft Fabric, including its discover and build cohorts capability, and features from the Fabric platform. Microsoft Fabric: is a unified, AI-powered data platform designed to simplify data management and analytics. It integrates various tools and services to handle every stage of the data lifecycle, including ingestion, preparation, storage, analysis, and visualization. Fabric is built on a Software as a Service (SaaS) foundation, offering seamless experience for organizations to make data-driven decisions. For additional details, refer to the following link: What is Microsoft Fabric - Microsoft Fabric | Microsoft Learn Healthcare data solutions in Fabric: Healthcare data solutions in Fabric help you accelerate time to value by addressing the critical need to efficiently transform healthcare data into a suitable format for analysis. With these solutions, you can conduct exploratory analysis, run large-scale analytics, and power generative AI with your healthcare data. By using intuitive tools such as data pipelines and transformations, you can easily navigate and process complex datasets, overcoming the inherent challenges associated with unstructured data formats. For additional details, refer to the following links: Healthcare data solutions in Microsoft Fabric - Microsoft Cloud for Healthcare | Microsoft Learn Discover and build cohorts: Discover and build cohorts (preview) capability in healthcare data solutions enables healthcare organizations to efficiently analyze and query healthcare data from multiple sources and formats. It simplifies the preparation of data for health trend studies, clinical trials, quality assessments, historical research, and AI development. It supports natural language queries for multimodal data exploration and cohort building, making it ideal for research and AI-driven projects. For additional details, refer to the following link: Overview of discover and build cohorts (preview) - Microsoft Cloud for Healthcare | Microsoft Learn The proposal for research data platform architecture builds upon the following foundational premises: Recognition of Fabric as the all-in-one data storage, processing, management and analytics platform with enterprise-level features around security, availability and self-service. Adoption of Fabric Workspace(s) as the security boundary (a secure logical container) for maintaining data platform items (data storage and processing assets). Fabric workspaces may be provisioned for and used by different research data platform stakeholders (groups of users) with different requirements around use cases, data privacy, data sensitivity and access security. Use of healthcare data solutions in Fabric, as the core capability to maintain healthcare data assets in a standard (interoperable) manner. Healthcare data solutions enables the storage and processing of several healthcare data modalities and formats that follow industry standards (for example, clinical modality in FHIR® NDJSON format and Clinical-Imaging modality’s DICOM® format). Industry standards make it easier for research data platform stakeholders to share (exchange) data and insights within their own organization as well as (when needed) with other organizations that they collaborate with. Use of Fabric native capabilities to address requirements that may not (yet) have been implemented for healthcare specific needs. This provides the research data platform stakeholders with the flexibility to develop various data storage and processing workloads easily in a low (or no) code manner. Fig – Conceptual architecture of research data platform in Microsoft Fabric Note: This diagram is an architectural pattern and does not constitute one to one mapping of existing Microsoft products. Organizing source data in data workspace (One Data Hub in the above diagram) Organize your enterprise data into a data workspace that could be leveraged for research purposes. This acts as a ‘One Data Hub’ for the research data platform. Multiple Lakehouse can be present in this workspace. There should be at least one Lakehouse that organizes data using ‘unified folder structure’ best practice. Convert data from non-supported format to healthcare data solutions supported format to leverage out of the box transformation for multimodal data: For healthcare data solutions supported modalities: Implement custom transformations to convert data to supported modalities/format. For unsupported modalities: Implement extensions to bronze Lakehouse to accommodate additional data modalities. Epic data availability: Epic supports FHIR data export using Bulk FHIR APIs. If your dataset meets the use cases of Epic Bulk Data, you can store the resulting FHIR resources into One Data Hub for further transformation. Avoid data content duplication: Data duplication cannot be totally avoided. However, the same file and same content are never duplicated. There will be situations when data needs to be transformed to suit the needs of existing transformation pipelines for accelerating research data platform development. Additionally, OneLake in Fabric storage, where Lakehouse is maintained, uses file compression. Healthcare data solutions in Fabric has functionality to compress raw files to zip and always writes structured data to delta parquet which is a higher compressed format. More information can be found here - Data architecture and management in healthcare data solutions - Microsoft Cloud for Healthcare | Microsoft Learn Curating data for research (One Analytics workspace in the above diagram) Implement and extend Silver Lakehouse: A flattened FHIR® data model is provided by healthcare data solutions out of the box within the Silver Lakehouse. Extending the existing data model is possible through adding new columns to existing tables or through adding new tables in the Silver Lakehouse. If there is a need to introduce a different data model altogether, it is best to implement it using a different Lakehouse. Implement and extend Gold Lakehouse: Deploy and extend Observational Medical Outcomes Partnership Common Data Model (OMOP CDM): Deploy OMOP CDM 5.4 out of the box with healthcare data solutions deployment. Extend OMOP CDM to accommodate additional modalities. For example, implement Gene sequencing, Variant occurrence and Variant annotation tables to add genomics modality into OMOP CDM or implement medical imaging data on OMOP CDM as described here - Development of Medical Imaging Data Standardization for Imaging-Based Observational Research: OMOP Common Data Model Extension - PubMed Implement custom Gold Lakehouse(s): Implement other custom Gold Lakehouse using Fabric tools that run your transformation logic from Silver to Gold. These Lakehouse cannot be connected to discover and build cohorts capability within healthcare data solutions. Customers that need access to custom data can connect their custom cohort browsers to the SQL Analytics Endpoint(s) of their custom Gold Lakehouse(s). Enable data de-identification: Microsoft provides several solutions that can be used to implement a comprehensive de-identification solution that customers expect. Refer to the articles below for details. Dynamic data masking in Fabric Data Warehouse - Microsoft Fabric | Microsoft Learn Row-level security in Fabric data warehousing - Microsoft Fabric | Microsoft Learn Column-level security in Fabric data warehousing - Microsoft Fabric | Microsoft Learn Announcing a de-identification service for Health and Life Sciences | Microsoft Community Hub Cohort discovery using cohort builder tool Microsoft’s cohort browser: Today Discovery and Build Cohort supports eyes-on cohort discovery. This is an out of the box solution that is part of healthcare data solutions in Fabric. When eyes off discovery is supported, researchers as well as research IT can benefit from both eyes off and eyes on discovery and cohort building. 3rd-party cohort browser (e.g., OHDSI Atlas): Most 3rd party cohort browsers (E.g. OHDSI Atlas) and home-grown cohort browsers typically support connection to a SQL endpoint. Microsoft Fabric platform provides the capability of exposing SQL endpoint from a Lakehouse that can be connected to a 3rd party cohort browser to perform cohort discovery. Automated data delivery Creating research workspaces with cohort needed for research: Create separate workspaces for different research projects to keep Fabric items distinct and project specific using Fabric APIs. Assign workspaces to a Fabric capacity: Note: When needed, and if the organization has more than one Fabric capacity provisioned, workspace assignment can be spread across different capacities to help manage cost and performance. Next, set up a Lakehouse and provide access for team members (as per IRB approval list). This ensures both access and security at the workspace level. Export data to research workspace (format desired by researchers): Currently, DBC exports data as CSV/JSON files stored in a Lakehouse within the same workspace. Shortcut the destination Lakehouse into research workspace to keep the sanity of cohort data. Tools for researchers: Fabric provides several data engineering and data science tools out of the box that researchers can leverage to perform research. The following are some of the documents that customers can use to enable researchers with the tools of choice. Data science in Microsoft Fabric - Microsoft Fabric | Microsoft Learn Create, configure, and use an environment in Fabric - Microsoft Fabric | Microsoft Learn Migrate libraries and properties to a default environment - Microsoft Fabric | Microsoft Learn Charge back: Fabric compute pricing depends on the chosen Fabric capacity SKU. Assigning different Fabric capacities to different projects or groups within the same cost center can facilitate chargeback. See the step mentioned above on assigning a workspace to a Fabric capacity during workspace creation. Manage historic data migration to the research data platform on Fabric In most instances, customers already possess a research data platform. They seek to transition to this proposed solution without disrupting their current research data flow and obligations. Follow this approach to migrate or use data from the existing platform to the new one: Use your current research data platform as a Lakehouse or a Data Warehouse in Fabric (take advantage of Shortcut and Mirroring features available in Fabric). Fabric offers cross-database query, i.e. allowing to query and join multiple Lakehouse and data warehouses in a single query. Customers can choose how and where to implement such queries to augment the healthcare data solutions datasets with their existing datasets, all natively in Fabric. A bridge/mapping layer can be built to link the old and the new in a cross-relational way. Conceptually, such an approach has also ties to Bring Your Own Database (BYO-DB) requirement, which is the ability to bring custom defined format and still be able to easily convert to healthcare data solutions specific format. Other workflow integration Integrate research data platform with IRB workflow: IRB workflows are dependent on the tools utilized. For instance, eIRB solution from Huron. While there is currently no direct integration between IRB workflows and the research data platform on Fabric, it is possible to develop a connector using Power Platform integration with Fabric. Specific details are not available at this time as this remains an exploratory initiative. Another approach will be to use Fabric REST APIs (as a pro-code method) that can enable richer integration between Fabric and the 3 rd -party system, and a better consuming user experience at the end. Capture logs necessary for “accounting of disclosures”: Logs in Fabric can be captured at event level. It’s up to the customer to decide the level and type of logs that need to be captured for accounting of disclosure. This will need some custom implementation. One such capability of Fabric that can be used is: Track user activities in Microsoft Fabric - Microsoft Fabric | Microsoft Learn FHIR® is a registered trademark of Health Level Seven International, registered in the U.S. Trademark Office and is used with their permission. DICOM® is the registered trademark of the National Electrical Manufacturers Association (NEMA) for its Standards publications relating to digital communications of medical information. If you are a Microsoft customer needing further information, support, or guidance related to the content in this blog, we recommend you reach out to your Microsoft account team in order to set up a discussion with the authors.
Fabric Data Agents: Unlocking the Power of Agents as a Steppingstone for a Modern Data Platform
What Are Fabric Data Agents? Fabric Data Agents are intelligent, AI-powered assistants embedded within Microsoft Fabric, a unified data platform that integrates data ingestion, processing, transformation, and analytics. These agents act as intermediaries between users and data, enabling seamless interaction through natural language queries in the form of Q&A applications. Whether it's retrieving insights, analyzing trends, or generating visualizations, Fabric Data Agents simplify complex data tasks, making advanced analytics accessible to everyone—from data scientists to business analysts to executive teams. How Do They Work? At the center of Fabric Data Agents is OneLake, a unified and governed data lake that joins data from various sources, including on-premises systems, cloud platforms, and third-party databases. OneLake ensures that all data is stored in a common, open format, simplifying data management and enabling agents to access a comprehensive view of the organization's data. Through Fabric’s Data Ingestion capabilities, such as Fabric Data Factory, OneLake Shortcuts, and Fabric Database Mirroring, Fabric Data Agents are designed to connect with over 200 data sources, ensuring seamless integration across an organization's data estate. This connectivity allows them to pull data from diverse systems and provide a unified analytics experience. Here's how Fabric Data Agents work: Natural Language Processing: Using advanced NLP techniques, Fabric Data Agents enable users to interact with data through conversational queries. For example, users can ask questions like, "What are the top-performing investment portfolios this quarter?" and receive precise answers, grounded on enterprise data. AI-powered Insights: The agents process queries, reason over data, and deliver actionable insights, using Azure OpenAI models, all while maintaining data security and compliance. Customization: Fabric data agents are highly customizable. Users can provide custom instructions and examples to tailor their behavior to specific scenarios. Fabric Data Agents allow users to provide example SQL queries, which can be used to influence the agent’s behavior. They also can integrate with Azure AI Agent Service or Microsoft Copilot Studio, where organizations can tailor agents to specific use cases, such as risk assessment or fraud detection. Security and Compliance: Fabric Data Agents are built with enterprise-grade security features, including inheriting Identity Passthrough/On-Behalf-Of (OBO) authentication. This ensures that business users only access data they are authorized to view, keeping strict compliance with regulations like GDPR and CCPA across geographies and user roles. Integration with Azure: Fabric Data Agents are deeply integrated with Azure services, such as Azure AI Agent Service and Azure OpenAI Service. Practically, organizations can publish Fabric Data Agents to custom Copilots using these services and use the APIs in various custom AI applications. This integration ensures scalability, high availability, and performance and exceptional customer experience. Why Should Financial Services Companies Use Fabric Data Agents? The financial services industry faces unique challenges, including stringent regulatory requirements, the need for real-time decision-making, and empowering users to interact with an AI application in a Q&A fashion over enterprise data. Fabric Data Agents address these challenges head-on through: Enhanced Efficiency: Automate repetitive tasks, freeing up valuable time for employees to focus on strategic initiatives. Improved Compliance: Use robust data governance features to ensure compliance with regulations like GDPR and CCPA. Data-Driven Decisions: Gain deeper insights into customer behavior, market trends, and operational performance. Scalability: Seamlessly scale analytics capabilities to meet the demands of a growing organization, without really investing in building custom AI applications which require deep expertise. Integration with Azure: Fabric Data Agents are natively designed to integrate across Microsoft’s ecosystem, providing a comprehensive end-to-end solution for a Modern Data Platform. How different are Fabric Data Agents from Copilot Studio Agents? Fabric Data Agents and Copilot Studio Agents serve distinct purposes within Microsoft's ecosystem: Fabric Data Agents are tailored for data science workflows. They integrate AI capabilities to interact with organizational data, providing analytics insights. They focus on data processing and analysis using the medallion architecture (bronze, silver, and gold layers) and support integration with the Lakehouse, Data Warehouse, KQL Databases and Semantic Models. Copilot Studio Agents, on the other hand, are customizable AI-powered assistants designed for specific tasks. Built within Copilot Studio, they can connect to various enterprise data sources like OneLake, AI Search, SharePoint, OneDrive, and Dynamics 365. These agents are versatile, enabling businesses to automate workflows, analyze data, and provide contextual responses by using APIs and built-in connectors. What are the technical requirements for using Fabric Data Agents? A paid F64 or higher Fabric capacity resource Fabric data agent tenant settingsis enabled. Copilot tenant switchis enabled. Cross-geo processing for AIis enabled. Cross-geo storing for AIis enabled. At least one of these: Fabric Data Warehouse, Fabric Lakehouse, one or more Power BI semantic models, or a KQL database with data. Power BI semantic models via XMLA endpoints tenant switchis enabled for Power BI semantic model data sources. Final Thoughts In a data-driven world, Fabric Data Agents are poised to redefine how financial services organizations operate and innovate. By simplifying complex data processes, enabling actionable insights, and fostering collaboration across teams, these intelligent agents empower organizations to unlock the true potential of their data. Paired with the robust capabilities of Microsoft Fabric and Azure, financial institutions can confidently navigate industry challenges, drive growth, and deliver superior customer experiences. Adopting Fabric Data Agents is not just an upgrade—it's a transformative step towards building a resilient and future-ready business. The time to embrace the data revolution is now. Learn how to create Fabric Data AgentsAnnouncing Mirroring for Azure Database for PostgreSQL in Microsoft Fabric for Public Preview
Back at the first European Microsoft Fabric Community Conference in September 2024 we announced our Private Preview program for Mirroring for Azure Database for PostgreSQL in Microsoft Fabric. Today, in conjunction with 2025 edition of Microsoft Fabric Community Conference in Las Vegas, we're thrilled to announce our Public Preview milestone, giving customers the ability to leverage friction-free near-real time replication from Azure Database for PostgreSQL flexible server to Fabric OneLake in Delta tables, providing a solid foundation for reporting, advanced analytics, AI, and data science on operational data with minimal effort and impact on transactional workloads. Mirroring is setup from Fabric Data Warehousing experience by providing the Azure Database for PostgreSQL flexible server and database connection details, provide selections on what needs to be mirrored into Fabric, either all data or user selected eligible mirrored tables. And, just like that, mirroring is ready to go. Mirroring Azure Database for PostgreSQL flexible server creates an initial snapshot in Fabric OneLake, after which data is kept in sync in near-real time with every transaction. How mirroring to Fabric works in Azure Database for PostgreSQL flexible server Fabric mirroring in Azure Database for PostgreSQL flexible server is based on principles such as logical replication and the Change Data Capture (CDC) design pattern. Once Fabric mirroring is established for a database in Azure Database for PostgreSQL flexible server, an initial snapshot is created by a background process for selected tables to be mirrored. That snapshot is shipped to a Fabric OneLake's landing zone in Parquet format. A process running in Fabric, known as replicator, takes these initial snapshot files and creates tables in Delta format in the Mirrored database artifact. Subsequent changes applied to selected tables are also captured in the source database and shipped to the OneLake landing zone in batches. Those batches of changes are finally applied to the respective Delta tables in the Mirrored database artifact. For Fabric mirroring, the CDC pattern is implemented in a proprietary PostgreSQL extension called azure_cdc, which is installed and registered in source databases during Fabric mirroring enablement workflow. This guided process has a new dedicated page in Azure Portal and is setting up all required pre-requisites and is offering a simplified experience where you just need to select which databases you want to replicate to Fabric OneLake (default is up to 3). You can read additional details regarding the server enablement process and other critical configuration and monitoring options on a dedicated page in Azure Database for PostgreSQL flexible server product documentation. Explore advanced analytics and data engineering for PostgreSQL in Microsoft Fabric Once data is on OneLake, mirrored data in the delta format is ready for immediate consumption across all Fabric experiences and features, such as Power BI with new Direct Lake mode, Data Warehouse, Data Engineering, Lakehouse, KQL Database, Notebooks and Copilot, which work instantly. Direct Lake mode is a fast path to load the data from the lake with groundbreaking semantic model capability for analyzing very large data volumes in Power BI. As Direct Lake mode also supports reading Delta tables right from OneLake, the Mirrored PostgreSQL database is Power BI ready along with Copilot capabilities. Data across any mirrored database (either Azure Database for PostgreSQL, Azure SQL DB, Azure Cosmos DB or Snowflake) can be cross-joined as well, enabling querying across any database, warehouse or Lakehouse (either as a shortcut to AWS S3 or ADLS Gen 2 etc.). With the same approach, you can also have multiple PosgreSQL databases from multiple servers mirrored to OneLake like in a typical SaaS provider scenario, where each database belongs to a different tenant, and execute cross-database queries to aggregate and analyze critical business metrics. Data scientists and data engineers can work with the mirrored Azure Database for PostgreSQL data joined with other sources (see this example with CosmosDB data) that are created as shortcuts in Lakehouse. Read about endless possibilities when loading operational databases in OneLake and Microsoft Fabric in related section of our product documentation here. Getting started with Mirroring for Azure Database for PostgreSQL in Fabric To summarize, Mirroring Azure Database for PostgreSQL in Microsoft Fabric plays a crucial role in enabling analytics and driving insights from operational data by ensuring that the most recent data is available for analysis. This allows businesses to make decisions based on the most current situation, rather than relying on outdated information. Improving accuracy also reduces the risk of discrepancies between the source and the replicated data, leading to more accurate analytics and reliable insights. In addition, is essential for predictive analytics and AI models provide the most recent data to make accurate predictions and decisions. To get started and learn more about Mirroring Azure Database for PostgreSQL flexible server in Microsoft Fabric, its pre-requisites, setup, FAQ’s, current limitations, and tutorial, please click here to read all about it and stay tuned for more updates and new features coming soon. To get more updates also on overall Mirroring capabilities in Fabric, please read this other blog post where you will get the latest news.
