Month: May 2026

🔹 Feature: Bicep CLI — and Especially the Bicep Console 🔹 What It Does: The Bicep CLI is so much more than just a compiler to ARM JSON. It builds, lints, restores, decompiles, formats, snapshots, and offers JSON-RPC — but the standout feature is bicep console, a fully interactive REPL where you can type Bicep expressions and instantly see the results. A safe, local playground for your infrastructure code. 🧪 What Is It Giving You: ✅ Instant Expression Evaluation: Type any valid Bicep expression and see the output immediately — no build, no deploy, no Azure round-trip. Learning a new function? Test it live. ✅ Variables + Multi-Line Support: Declare variables using var and reuse them. Multi-line complex expressions are supported with automatic structural detection. ✅ Real-Time Syntax Highlighting: Input and output are colorized so you catch mistakes before they reach a template. Supports nearly all Bicep functions. ✅ Unit-Test Your User-Defined Functions: Validate your UDFs return exactly what you intend — no deployment required. Prototype, verify, and then commit. ✅ Validate AI-Generated Code Safely: Paste Copilot-generated Bicep functions into the console to confirm correctness before trusting them in production. ✅ Automation-Ready stdin/stdout: Since v0.40.2, piping input and output is … Bicep CLI — and Especially the Bicep ConsoleRead more

🔹 Feature: Bicep What-If — The “terraform plan” Equivalent for Azure Native IaC 🔹 What It Does: Love the safety of running Terraform plan before you apply? Bicep has the same thing. The what-if operation previews exactly how your resources will change before you deploy — Create, Modify, Delete — without touching a single live resource. No surprises, no “wait, why did that subnet disappear.” 🔍 What Is It Giving You: ✅ Dry-Run Before Every Deploy: what-if predicts the changes without making any. It diffs your Bicep file against the current state of Azure and shows you the delta. Just like a plan — but native, no extra tooling, no state file to babysit. ✅ Color-Coded Change Symbols: Output marks every change clearly — “+” Create, “~” Modify, “-” Delete. At a glance you see that a tag got dropped, an address prefix changed, or a subnet is about to be deleted. Read the diff, then decide. ✅ Confirm-With-What-If in One Command: Run az deployment group create –confirm-with-what-if (or -c) and Azure shows you the preview, then prompts “Are you sure?” Plan and apply in a single, gated step. This is the muscle memory Terraform people are looking for. ✅ … Bicep What-If — The “terraform plan” Equivalent for Azure Native IaCRead more

The big consumers inside each profile are usually stored in user profiles in subfolders, all of which are safe to delete (Chrome rebuilds them): To safely wipe caches across all profiles without losing logins, bookmarks, or passwords run this powershell script:

🔹 Feature: Video Recap in Microsoft 365 Copilot 🔹 What It Does: Copilot turns recorded Teams meetings into short, narrated video summaries — spotlighting the key moments, decisions, and discussion points. Catch up on a one-hour meeting in a couple of minutes, without scrubbing through the whole recording. 🎬 What Is It Giving You: ✅ Transcript-Driven Highlights: Copilot analyzes the meeting transcript, identifies the topics and decisions that mattered, and stitches relevant moments into a coherent recap that reflects the tone and flow of the original meeting. ✅ AI Narration + Captions: Each recap combines short clips (20–40 seconds) with generated AI narration and captions — delivering context and accessibility, not just a silent montage. ✅ One-Prompt Generation: In Copilot Chat, just type Recap /meeting_name and hit send. Copilot handles editing — no timeline scrubbing or manual clipping. ✅ Storage-Smart by Design: Recaps are generated from the original recording already in OneDrive or SharePoint. Copilot adds highlights, captions, and narration without duplicating the video. ✅ Catch Up Faster Across the Org: Missed the standup? Joined a project late? Video recap means onboarding into a discussion thread in minutes, not an afternoon of recordings. ⚠️ Requires a Microsoft 365 Copilot subscription … Video Recap in Microsoft 365 CopilotRead more

🔹 Feature: Microsoft Entra Authentication for Application Insights 🔹 What It Does: Ingest telemetry into Application Insights using Microsoft Entra ID instead of instrumentation keys. Managed identities replace shared secrets, and local authentication can be fully disabled — ensuring only Entra-authenticated telemetry reaches your resource. Critical for alerting, autoscaling, and all operational decisions built on telemetry you trust. 🔐 💡 What Is It Giving You: ✅ No More Instrumentation Keys in Config: Apps authenticate via their managed identity — no rotation, no secret leakage, no connection strings in pipelines, ARM templates, or Git. ✅ Disable Local Auth Entirely: Set DisableLocalAuth = true on your App Insights resource to reject API keys and legacy instrumentation keys. Pure Zero Trust telemetry ingestion. ✅ Built-in Azure Policy Enforcement: Use the built-in policy “Application Insights components should block non-Azure Active Directory based ingestion” to audit or enforce across subscriptions or management groups — no custom policies needed. ✅ Broad SDK Coverage — GA: ASP.NET Core, .NET Classic, Java Agent (stable), Node.js, Python — all support Entra auth via DefaultAzureCredential or ManagedIdentityCredential. ✅ Zero-Code Auto-Instrumentation: For Azure App Service, set APPLICATIONINSIGHTS_AUTHENTICATION_STRING=Authorization=AAD (system-assigned) or Authorization=AAD;ClientId={uami-client-id} (user-assigned). Done. ✅ Monitoring Metrics Publisher Role: A single built-in RBAC … Microsoft Entra Authentication for Application InsightsRead more

🔹 Feature: Azure Container Apps Express 🔹 What It Does: The fastest path from container image to internet-reachable app on Azure. No environment to provision, no networking to configure, no scaling rules to write. Bring an image, get a URL — done. Express runs on pre-provisioned consumption capacity with production-ready defaults baked in, while keeping ACA’s serverless model (scale-to-zero included). What Is It Giving You: ✅ Zero Infrastructure Decisions: No Container Apps environment to create or manage. No workload profile to size. No scaling rules to author. The platform allocates infrastructure, applies opinionated defaults, and gets your app online in minutes. ✅ Built-In Production Defaults: Ingress, secrets, environment variables, observability, and cold-start optimization are pre-configured. Same serverless engine as standard ACA — just without the setup tax. ✅ Purpose-Built for Two Audiences: Developers who want to ship fast (SaaS apps, APIs, web dashboards, prototypes) and AI agents that deploy on demand (MCP servers, tool-use endpoints, multi-step workflow APIs, human-in-the-loop UIs). Agent-first deployment is finally a first-class scenario. ✅ Consumption-Based, Scale-to-Zero: Pay-as-you-go CPU compute that runs only when requests arrive. Apps idle to zero between traffic — perfect for prototypes, demos, and bursty agent workloads. ✅ One-Command Deploy: az containerapp up … Azure Container Apps ExpressRead more

🔹 Feature: Azure Resource Manager MCP Server — Public Preview – Finally 🔹 What It Does: A remote MCP server that gives AI agents first-class access to Azure infrastructure operations through Azure Resource Manager. Generate, validate, and execute Azure Resource Graph queries from natural language, deploy and monitor ARM templates, and let your AI agent reason across your entire Azure estate — all secured by your existing RBAC and Azure Policy. 🤖 💡 Thesis: AI agents can now operate Azure infrastructure natively — no more screenshots, manual KQL, or brittle scripts. What Is It Giving You: ✅ Natural Language → Azure Resource Graph: Stop writing KQL from scratch. Tell your agent “Find all VMs in my subscription without managed disks” or “Show me resources created in the last 30 days without required tags” — the server generates the ARG query, validates it, and executes it across every Azure resource type. ✅ Full ARM Template Lifecycle: Pass an ARM template, kick off deployment at resource group scope, check deployment status, and cancel in-flight deployments if something looks wrong. AI-driven infra deployment with proper guardrails. ✅ Remote MCP — Not Yet Another Local Server: Unlike most MCP servers you spin up on … Azure Resource Manager MCP ServerRead more

🔹 Feature: Breakout Rooms in Microsoft Teams 🔹 What It Does: Turn a 300-person meeting from “everyone listening like radio” into real, productive working sessions — split the audience into parallel rooms where people actually engage instead of passively watching. 🆕 What It Gives You: ✅ Scale to 300 Participants: Run interactive sessions with audiences far beyond the typical 100-person ceiling — great for large workshops, training, or all-hands working sessions. ✅ Up to 50 Parallel Rooms: Split a single meeting into focused groups for discussions, exercises, design sprints, or domain-specific tracks. ✅ Pre-Meeting Setup: Create rooms and assign participants from the calendar invite before anyone joins — show up to a meeting that’s already organized. ✅ Smart Assignment Modes: Auto-assign, manual assignment, or let participants self-select rooms — flexibility for any session format. ✅ Shuffle Function: Randomize participants across rooms with one click — perfect for icebreakers, mixed-team exercises, or rotating breakouts. ✅ Time Limits: Set a countdown — rooms close automatically and everyone returns to the main meeting. No more “just five more minutes” drift. ✅ Broadcast Announcements: Send a message to all rooms at once — keep parallel groups synchronized without interrupting each session manually. ✅ Delegate … Breakout Rooms in Microsoft TeamsRead more

🔹 Feature: New Endpoint Security for Azure Virtual Desktop & Windows 365 🔹 What It Does: Stop keyloggers from capturing what you type, and stop screen scrapers from capturing what you see. 🆕 What’s New: ✅ Windows Cloud Keyboard Input Protection: Establishes a secure channel from the endpoint kernel all the way to the Cloud PC or AVD session host. Keystrokes are encrypted at the kernel driver level on the physical device and only decrypted inside the remote VM — keyloggers and keystroke injection malware on the endpoint see nothing but ciphertext. ✅ Screen Capture Protection: Blocks screen recording tools, screenshot utilities, and screen-sharing apps from capturing AVD/W365 session content. Available in two modes — block client only, or block client + everything in front of it (camera-of-screen scenarios still need DLP). ✅ BYOD-Ready Trust Boundary: Shifts the security perimeter from “trust the device” to “trust the encrypted channel” — critical when users connect from unmanaged Windows 11 personal devices. ✅ Kernel-Level Enforcement: Software kernel driver + system-level encryption service routes all keyboard input directly to the VM in encrypted format — bypasses user-mode interception entirely. ✅ Transparent to Users: No workflow change, no perceptible latency, no IT admin overhead … New Endpoint Security for Azure Virtual Desktop & Windows 365Read more

🔹 Feature: SQL Database in Microsoft Fabric 🔹 What It Does: A fully managed, cloud-native SQL engine inside Fabric — built for operational workloads, AI apps, and the data plumbing that powers analytics. 💡 Where teams actually use it: Meta-driven pipelines Power BI writebacks Reference data store Operational data store (ODS) Line of business Power Apps + Translytical Vector and AI app dev Event-driven apps Reverse ETL 🆕 What’s New: ✅ Migration Experience in Fabric: Move existing SQL workloads into Fabric with a guided in-portal experience — no more manual schema/data juggling. schema/data juggling. ✅ Point-in-Time Restore for Dropped Databases: Accidentally deleted database? Recover it to any point in time within retention — safety net built in. ✅ List Deleted Databases by Workspace: Full visibility of what was dropped and when — no more guessing during recovery. ✅Pre/Post Deployment Actions: Hook custom logic into your SQL project deployments — seed data, set permissions, run validations automatically. ✅ Custom Table Selection for Mirroring: Pick exactly which tables get mirrored to OneLake — control cost, scope, and refresh patterns per workload. ✅ SQL Projects Open in SSMS: Edit and manage Fabric SQL projects directly from SSMS — the tool every DBA already … SQL Database in Microsoft FabricRead more

🔹 Feature: Extended SharePoint Permissions (ESP) 🔹 What It Does: Your SharePoint permissions — extended everywhere. Lose access in SharePoint, and the downloaded file won’t open anymore. Anywhere. 💡 Thesis: The perimeter follows the file, not the other way around. What is it giving you: ✅ Permissions That Travel: SharePoint site permissions are automatically applied to files when downloaded, copied, or moved off the site — no manual labeling required. ✅ Just-in-Time Protection: If permissions are revoked, the file is deleted, the site goes inactive, or the file is moved — the downloaded copy stops opening. Instantly. ✅ Live Permission Sync: Change a user’s SharePoint access and the change reflects immediately on every downloaded copy of the file. ✅ Move/Copy Lockdown: Files can’t be moved or copied to a different site. Within the same site only if the user has list create/delete rights. ✅ Zero-Effort Rollout: Applies to unlabeled files and files with non-encrypting labels — perfect for organizations early in their labeling journey. ✅ Permission Mapping Built-In: SharePoint Owner/Edit/Read maps directly to RMS usage rights (Owner/Editor/Viewer) — no custom rights policies to design. ⚠️ Worth knowing: Requires Microsoft 365 Apps 2402+ (Current/Monthly Enterprise/Semi-Annual) Files won’t open offline — connection … Extended SharePoint Permissions (ESP)Read more

🔹 Feature: Foundry Local (Orion) Edge AI SDK 🔹 What It Does: Run AI models directly on users’ devices without any separate runtime installation, setup scripts, or external dependencies. 💡 Thesis: Edge AI adoption accelerates when it’s truly plug-and-play. No more winget install foundry What it gives you: ✅ Zero-Setup Deployment: Your app includes the AI runtime out of the box—no need for users to install extra components. ✅ On-Device Execution: Models execute locally on CPU, GPU, or NPU for full privacy, offline functionality, and no incremental cloud cost. ✅ Seamless API Integration: Compatible with existing REST and gRPC APIs—just point your endpoints and keep your code intact. ✅ Adaptive Performance: Automatically detects available hardware and picks the optimal model variant for the user’s machine. ✅ Production-Grade Reliability: Powered by the same engine used in enterprise edge workloads; robust in both desktop and embedded scenarios. ✅ Developer-First Experience: SDKs for Python, C++, JavaScript, and C#—drop in and start shipping AI capabilities immediately. 🚀 The shift: Stop making users install AI stacks. Deliver AI as a self-contained part of your application. 🌐 https://learn.microsoft.com/en-us/azure/foundry-local/concepts/foundry-local-architecture

🔹 Feature: azd ai agent extension 🔹 What It Does: Scaffold, provision, and deploy AI agents to Microsoft Foundry — straight from your terminal, from code to live endpoint in minutes. What is it giving you: ✅ End-to-End Lifecycle: From init through provisioning, deployment, monitoring, and CI/CD — one consistent CLI experience. ✅ Declarative Config: azure.yaml and agent.yaml define services, models, and dependencies — all version-controlled with your code. ✅ One-Command Deploy: azd up builds the container, pushes to ACR, provisions Foundry resources, deploys models, and registers the agent. ✅ Local Dev Loop: azd ai agent run runs the agent locally against your remote Azure resources for fast iteration. ✅ Pipeline Ready: Generate Azure DevOps or GitHub Actions pipelines directly from the CLI — production path from day one. 🚀 Fast run: azd init azd up azd monitor azd pipeline config 🌐 https://learn.microsoft.com/en-us/azure/developer/azure-developer-cli/extensions/azure-ai-foundry-extension 🌐 https://devblogs.microsoft.com/azure-sdk/azure-developer-cli-foundry-agent-extension/

🔹 Feature: Hosted Agents in Foundry Agent Service 🔹 What It Does: Run your containerized AI agents on Microsoft-managed, agent-optimized infrastructure — simply bring your code, and Foundry handles the rest. What Is It Giving You: ✅ Bring Your Own Code: Package any framework — Microsoft Agent Framework, LangGraph, Claude Agent SDK, or OpenAI Agents SDK — into a container and deploy effortlessly. ✅ Per-Session Sandboxes: Enjoy hypervisor-level isolation with a persistent filesystem ($HOME and /files), allowing stateful session resumes. ✅ Scale to Zero: Pay nothing while idle; cold starts are predictable when sessions spin up again. ✅ Dedicated Agent Identity: Each agent receives its own Entra Agent ID and endpoint, ensuring no shared service accounts and a full audit trail. ✅ Enterprise Ready: Includes VNet integration, DLP policies, Responsible AI guardrails, OpenTelemetry-based observability, and built-in AI Red Teaming. ✅ Multi-Protocol: Expose Responses and/or Invocations endpoints from the same container to integrate seamlessly with Copilot, Teams, or custom applications. 🌐 https://learn.microsoft.com/en-us/azure/foundry/agents/concepts/hosted-agents 🌐 https://learn.microsoft.com/en-us/azure/foundry/agents/quickstarts/quickstart-hosted-agent 🌐 https://devblogs.microsoft.com/foundry/introducing-the-new-hosted-agents-in-foundry-agent-service-secure-scalable-compute-built-for-agents/

🔹 Feature: Managed Identity on Azure Arc-enabled Servers 🔹 What It Does: Hybrid and on-prem servers connected via Azure Arc can now obtain a system-assigned managed identity from Microsoft Entra ID — and use it to authenticate to Azure resources like Key Vault, Storage, or any ARM API. No client secrets baked into config files. No manual service account rotations. Just a token on demand via the local IMDS endpoint. 🔐 What Is It Giving You: ✅ Zero Stored Credentials: Workloads on your on-prem or multicloud servers authenticate to Azure without storing passwords, certificates, or service principal secrets locally. Microsoft Entra ID issues short-lived tokens via the local IMDS endpoint — just like Azure VMs. ✅ Identical Pattern as Azure VMs: The IMDS_ENDPOINT (http://localhost:40342) and IDENTITY_ENDPOINT environment variables are populated automatically. Code that already uses managed identities on Azure VMs works on Arc-enabled servers with zero changes. ✅ Hardened Token Acquisition: Token requests use a challenge-response flow — only higher-privileged users or groups (Admins / himds) can read the secret file to obtain tokens. Random processes can’t just grab tokens. ✅ Direct Key Vault Integration: Pull TLS certificates, API keys, and connection strings into IIS, NGINX, or any app without … Managed Identity on Azure Arc-enabled ServersRead more

🔹 Feature: Real-Time Intelligence in Microsoft Fabric 🔹 What It Does: Provides an end-to-end streaming analytics service that ingests, transforms, stores, queries, visualizes, and reacts to data in motion — all without writing any orchestration code. Perfect for IoT telemetry, CDC streams, logs, or fraud detection. What Is It Giving You: ✅ Real-Time Hub as a Central Catalog One unified place to discover, share, and govern all data streams in your organization — from Event Hubs, IoT Hub, Azure SQL CDC, Cosmos DB CDC, PostgreSQL CDC, Fabric workspace events, and Azure Blob Storage events. Streams become first-class citizens. ✅ Eventstreams — No-Code Stream Processing Connect to Apache Kafka, Kinesis, GCP Pub/Sub, MQTT, and database change feeds. Route events with filtering, transformations, windowed aggregations, and duplicate detection — no manual Flink or Spark setup required. ✅ Eventhouses + KQL Store and query structured, semi-structured, or unstructured streaming events in auto-partitioned, time-indexed storage. Enjoy lightning-fast queries with KQL, T-SQL, or natural language through Kusto Copilot. ✅ Digital Twin Builder Model your physical environment as an ontology, mapping assets and semantic relationships. Analyze time-series data spanning months to provide actionable context for operational decision-makers. ✅ Map for Geospatial Intelligence Visualize real-time data as … Real-Time Intelligence in Microsoft FabricRead more