Browse Artificial Intelligence Community (148)

In this community deep dive, junjieli walks through the GA release of Microsoft Foundry Toolkit for Visual Studio Code—covering model experimentation, agent development (no-code and code-first), evaluations, deployment to Microsoft Foundry Agent Service, and workflows for converting, profiling, and fine-tuning local models on Windows.

carlottacaste spotlights Athiq Ahmed’s winning Agents League Reasoning Agents project, CertPrep, detailing a Microsoft Foundry-based multi-agent pipeline that builds study plans, tracks readiness, generates assessments, and applies guardrails and human approval steps.

Sreekanth_Thirthala announces a public preview feature for Azure API Center: a plugin marketplace endpoint that lets developers discover and install AI plugins (including MCP servers and skills) from tools like Claude Code and GitHub Copilot CLI, while keeping enterprise governance and auth intact.

chandanAggarwal announces the public preview of Container Network Insights Agent, an agentic AI assistant for diagnosing AKS networking issues by correlating Kubernetes, Cilium/Hubble, and node-level Linux telemetry, then producing evidence-backed root-cause reports and remediation commands.

PrabhKaur (co-authored with Avneesh Kaushik) lays out an architecture-focused checklist for building AI agents in Microsoft Foundry with security, observability, least privilege, continuous validation, and human accountability built in from the start.

simonjj shares an Azure Developer CLI template that deploys Google’s Gemma 4 (via Ollama) onto Azure Container Apps serverless GPU with an OpenAI-compatible endpoint, protected by an Nginx basic-auth proxy, plus steps to verify the API and wire it into the OpenCode terminal coding agent for private, in-subscription prompting.

sachoudhury explains GitHub Copilot Custom Skills: repo- or user-scoped SKILL.md “runbooks” that Copilot can discover and execute in agent mode to automate multi-step developer workflows (commands, scripts, and report generation).

ManishChopra outlines six practical integration patterns for building agents and copilots that query Oracle Database@Azure with sub-millisecond proximity to Microsoft’s AI stack, covering options from Copilot Studio connectors to ORDS/PL/SQL, Azure Functions, and Logic Apps, plus the identity/governance controls typically needed for production.

jordanselig shows how to add runtime governance to a multi-agent ASP.NET Core travel planner on Azure App Service using the Microsoft Agent Governance Toolkit, including YAML policy allowlists, audit logging into Application Insights, and SRE controls like SLOs and circuit breakers.

mosiddi explains how Microsoft’s open-source Agent Governance Toolkit implements production-grade security and reliability controls for autonomous AI agents, covering its package architecture, policy enforcement (Agent OS), zero-trust identity (Agent Mesh), privilege rings (Agent Hypervisor), and SRE/observability integrations, including Azure deployment patterns.

jordanselig shows how to instrument Microsoft Agent Framework agents with OpenTelemetry GenAI semantic conventions and send that telemetry to Azure Application Insights, enabling the Agents (Preview) view for per-agent token usage, latency, errors, and end-to-end agent runs across an ASP.NET Core API and a WebJob.

jordanselig walks through building an MCP App (a tool plus a UI resource) with ASP.NET Core, rendering an interactive weather widget inside chat clients like VS Code Copilot, and deploying the MCP server to Azure App Service using azd and Bicep.

Shamir_AbdulAziz describes how Microsoft built Azure SRE Agent—an AI-powered ops agent—using “agentic workflows” across the SDLC, with human-in-the-loop governance, RBAC guardrails, and deep integration into telemetry and incident systems to reduce on-call toil and speed up incident mitigation.

Lee_Stott walks through what Azure Developer CLI (azd) is, why it’s useful for beginners, and how the AZD for Beginners learning path helps you move from local code to a repeatable Azure deployment workflow with templates, infrastructure as code, and lifecycle cleanup.

AnjaliSadhukhan argues that AI agents fail on enterprise questions mainly due to fragmented data and missing semantics, and outlines how Microsoft Fabric (OneLake, semantic models, Data Agents) and Azure AI Foundry can work together to provide governed, agent-ready access to business data.

Gaurav-Seth describes a hands-on, AI-guided workflow for migrating legacy IIS-hosted ASP.NET Framework apps to Managed Instance on Azure App Service, including how registry, storage, SMTP/MSMQ, and COM dependencies are handled via ARM templates and an install.ps1 startup script.

deepthihr walks through a real production incident running a private, enterprise AI platform on Azure, showing how DNS and private networking gaps (custom DNS, Private Endpoints, and Azure Container Apps internal ingress) caused intermittent failures—and the concrete fixes that stabilized the environment.

Pamela_Fox walks through implementing Model Context Protocol (MCP) server authentication with Microsoft Entra ID using the pre-registered (pre-authorized client) path, including Entra app registration setup, token validation in FastMCP, and an optional on-behalf-of flow to call Microsoft Graph securely.

macalde shares the March 2026 Innovation Challenge results, highlighting hackathon winners and example projects focused on building AI solutions on Azure (including RAG, multi-agent analytics, and governed AI outputs).

ShivaniThadiyan explains how Azure SQL Managed Instance is evolving from a SQL Server-compatible PaaS into an AI-enabled platform, covering built-in operational intelligence, vector search, in-database Python/R machine learning, and Copilot-assisted diagnostics with security and governance considerations.