How Thomson Reuters, Hexure, and CallRevu optimized their databases with Azure SQL Managed Instance

Because we're all about efficiency, why share one customer story when you can share three? Modernizing your data estate isn’t just a tech upgrade—it’s a business transformation. In our latest blog, we share how organizations like #ThomsonReuters, #Hexure, and #CallRevu are leveraging #Azure SQL Managed Instance to unlock scalability, security, and AI-driven innovation. ✔ Thomson Reuters streamlined tax prep for 7,000 businesses ✔ Hexure reduced processing times by 97% ✔ CallRevu improved lead conversion by 15% with real-time AI insights These stories are proving that modernization delivers measurable impact—faster operations, smarter decisions, and future-ready platforms.

#Azure AI Foundry Agent Service, #Model Context Protocol (#MCP) Server #Model Context Protocol (#MCP) Client In this module, I learned how to: ----Integrate external tools with Azure AI Foundry Agent Service using the Model Context Protocol (MCP). -----Set up an MCP server and client, and connect tools to an Azure AI Agent dynamically. AI agents are capable of performing a wide range of tasks, but many tasks still require them to interact with tools outside the large language model. Agents may need to access APIs, databases, or internal services. Manually integrating and maintaining these tools can quickly become complex, especially as your system grows, or changes frequently. Model Context Protocol (MCP) servers can help solve this problem by integrating with AI agents. Connecting an Azure AI Agent to a Model Context Protocol (MCP) server can provide your agent with a catalog of tools accessible on demand. This approach makes your AI solution more robust, scalable, and easier to maintain. Suppose you're working for a retailer that specializes in cosmetics. Your team wants to build an AI assistant that can help manage inventory by checking product stock levels and recent sales trends. Using an MCP server, you can connect the assistant to a set of tools that can make inventory assessments and provide recommendations to the team.

As efficiency becomes more important when using LLMs, model routing is an interesting and exciting solution for combining the power of multiple LLMs effectively. Check out Sanjeev's post on Azure Model Router to know more about the comprehensive router offering we have! Check out the Model Router events at the upcoming Microsoft Ignite!

🔴 That moment when you've triple-checked your Azure OpenAI API key, your endpoint looks perfect, and you STILL get a 401 Unauthorized error... Spent the last few days deep-diving into Azure AI Foundry Agents, and let me tell you: the authentication is a masterclass in subtle differences that will absolutely wreck your deployment. Here's what finally clicked: ❌ Azure OpenAI: `https://lnkd.in/gETC-pkU with `api-key` header ✅ Azure AI Foundry: `https://lnkd.in/gKhQQBqy with `api-key` header Same header. Different domain. Different API keys. The endpoint looks so similar you'll swear you have it right. But mixing Azure OpenAI credentials with AI Foundry endpoints (or vice versa) = instant 401. Even better? If local authentication is disabled on your Foundry resource, API keys won't work at all. You need Azure AD bearer tokens instead. After building both implementations side-by-side (Chat Completions vs. Agents API with RAG), I can now speak fluently in 401 error messages. 😅 🤔 Question for the Azure AI community: What's been your most frustrating authentication debugging experience? And did you eventually figure it out, or just pivot to a different approach? Building in public and documenting these landmines so others don't have to waste a day debugging the same issue. #AzureAI #AIFoundry #DeveloperLife #RAG #MachineLearning #Azure #TechnicalDebt

Thrilled to announce 🚀 Azure AI - Search Vector Stores support on LiteLLM (YC W23) - this brings direct querying of vector stores on Azure, to LiteLLM vs. the original Azure OpenAI integration which could only be referenced in Assistants API requests. (+4 more updates 👇) 💪 Vector Stores - view config.yaml vector store on UI 🛠️ Vector Stores - fix multiple values unpack error 💸 Responses API - support tag-based tracking 🎉 Responses API - restrict GET by team+user_id pair

🚨 Microsoft just revealed OpenAI lost $11.5 billion last quarter. To put that in perspective. Every 6 months OpenAI burns through the UK’s entire R&D budget. And it’s not a problem, it’s a strategy. The biggest subsidy in tech history? Microsoft isn’t “funding” OpenAI. It’s subsidising a moat. This is the price of entry to the next computing era and Microsoft knows it. While everyone debates profitability, Microsoft is quietly consolidating the AI stack: ➡️ Model layer (OpenAI) ➡️ Infrastructure (Azure GPUs) ➡️ Distribution (Copilot + Microsoft 365 + GitHub) By the time others can afford to compete, the market will already be consolidated. The moat will be built. This is what disruption looks like when capital is the barrier, not code.

💪 Empower your OneLake tables with graph-powered insights all in the Fabric Ecosystem: we are happy to announce the general availability of Neo4j Graph Intelligence for Microsoft Fabric! 0️⃣ Zero administration overhead 🫡 AI-assisted graph modeling 🫶 Seamless Fabric integration Find out more in this blog: https://bit.ly/3LnySey #Neo4j Microsoft Azure Stuart Moore

Neo4j Graph Intelligence for Microsoft Fabric is generally available now. Experience how easy it can be to transform tabular data into connected insights—whether you’re a business analyst exploring patterns or a data scientist building advanced pipelines.#Microsoft #Neo4j

💪 Empower your OneLake tables with graph-powered insights all in the Fabric Ecosystem: we are happy to announce the general availability of Neo4j Graph Intelligence for Microsoft Fabric! 0️⃣ Zero administration overhead 🫡 AI-assisted graph modeling 🫶 Seamless Fabric integration Find out more in this blog: https://bit.ly/3LnySey #Neo4j Microsoft Azure Stuart Moore

💪 Empower your OneLake tables with graph-powered insights all in the Fabric Ecosystem: we are happy to announce the general availability of Neo4j Graph Intelligence for Microsoft Fabric! 0️⃣ Zero administration overhead 🫡 AI-assisted graph modeling 🫶 Seamless Fabric integration Find out more in this blog: https://bit.ly/3LnySey #Neo4j Microsoft Azure Stuart Moore