Browse Artificial Intelligence Community (198)

nelsontam introduces Microsoft’s MARS (Map Autoregressive) model on Microsoft Foundry and explains how it converts satellite imagery into GIS-ready vector map data. The post outlines an end-to-end workflow with Microsoft Planetary Computer Pro, including STAC/COG ingestion requirements and practical deployment settings for running geospatial feature extraction at scale.

junjieli summarizes the July 2026 releases (1.6.3–1.6.6) of the Foundry Toolkit for VS Code, focusing on a flatter, tabbed workspace, inline tool wiring, deeper agent debugging via event-level inspection, and a new Agent Optimization preview that supports measured compare-and-deploy tuning for hosted agents.

KishoreKumarPattabiraman explains how to choose between building an interactive “skill” and an autonomous “sub-agent” when designing reusable AI capabilities, using practical checks around iteration style, voice fidelity, review gates, risk blast radius, and frequency of reuse.

Chris Noring shows how to build an AI shopping assistant that remembers customer preferences across sessions by plugging a SQL Server-backed history provider into Microsoft Agent Framework, then explains how the same setup can move from local Docker to Azure SQL Database (including Hyperscale) with only a connection-string change.

demiajayi introduces new Cost Management and Pricing toolsets for the Azure Resource Manager MCP server, enabling AI agents (including GitHub Copilot clients) to query costs, forecast spend, manage budgets and alerts, retrieve pricing, and analyze AKS costs directly through Azure’s control plane.

gauravbhardwaj explains how GitHub Copilot’s pooled included credits and metered overages work, and how to use user-level budgets and enterprise budgets to cap spend without unexpectedly blocking developers.

Aurnov Chattopadhyay recaps the June 2026 Azure Arc Server Forum, including updates on the Arc Server AI Agent, Azure Arc Multicloud Connector previews (GCP connector and EKS cluster enablement), and ESU guidance for Windows Server 2016/2012 and SQL Server 2016 end-of-support timelines.

Lee Stott walks through an end-to-end, production-minded approach for taking a multi-agent system from local development into Microsoft Foundry Agent Service, then distributing it to users in Microsoft Teams. The guide focuses on identity, governance, observability, and repeatable deployment using the FibreOps reference implementation.

Chris Noring explains how to investigate unexpected GitHub Copilot Enterprise AI-credit spend and then control it using GitHub Billing guardrails. The post walks through identifying the SKU driving costs, attributing usage to organizations and cost centers, and applying budgets, alerts, and per-user limits without breaking productive workflows.

prateekwrites explains why “cost per token” is a misleading metric for agentic systems, and proposes “cost per accepted task” as the unit that matters. The post maps this idea to an Azure reference architecture using APIM’s GenAI gateway, Azure AI Foundry evaluations, and a policy loop driven by telemetry and alerts.

Nora Zhan introduces “Physical-World Intelligence” and explains how GeoAI combines geospatial data (weather, satellite imagery, sensors, and maps) with enterprise context and AI. The article outlines Microsoft Foundry’s geospatial model catalog, Planetary Computer Pro as a GeoAI data plane, and a GeoAI SDK for production-scale inference workflows.

tonimontez explains how to govern GitHub Copilot’s usage-based spend using GitHub’s native budget controls first, then adds an Azure API Management gateway in front of an Azure AI Foundry (Azure OpenAI) deployment for real-time, token-granular quotas and per-developer telemetry.

vladvino announces the public preview of the AI Gateway tier for Azure API Management, focused on publishing and governing AI models and MCP servers. The post explains the new portal experience, policy-card governance (rate limits, quotas, Content Safety, fallback), and OpenTelemetry token metrics to destinations like Application Insights.

Lee Stott explains why the Model Context Protocol (MCP) is becoming the standard way AI agents connect to tools and data, then shows minimal runnable MCP server examples in Python and TypeScript plus practical guidance for VS Code hosting, security, and production operations.

carlottacaste shares a short walkthrough of the Foundry Toolkit for Visual Studio Code and a companion VS Code Learn course, focused on keeping model selection, prompt iteration, and agent development inside VS Code, with links to a full video playlist and a featured setup episode.

junjieli introduces the Foundry Agent Canvas (public preview), a GitHub Copilot App extension that lets developers discover resources, scaffold and configure a Foundry hosted agent, test it locally with an embedded Agent Inspector, and deploy it to Foundry Agent Service using azd-driven workflows.

brauerblogs shares a reminder to register for Microsoft’s “Path to Production for Agents” webinar series (July 27–28), focused on taking AI agent solutions from experimentation to secure, scalable production with guidance on governance, platform design, AgentOps, and multi-agent architecture patterns.

Kalaivanan explains how to use Azure API Management (APIM) as a control plane for Model Context Protocol (MCP) servers, focusing on enterprise-ready authentication, access control, observability, and governance. The post lays out practical patterns for putting APIM in front of MCP endpoints and using Entra ID, OAuth flows, and API Center for discovery.

richpaw describes a reference architecture for running Microsoft Discovery on a Windows VM in Azure and connecting it to an Azure CycleCloud HPC cluster via Azure NetApp Files (NFS) and SSH, so agentic workflows can submit Slurm jobs, read/write POSIX files, and stay inside a private network boundary.

richpaw explains how Microsoft Discovery combines agentic AI with Azure HPC to speed up semiconductor EDA workflows, focusing on closed-loop orchestration across tools, data, and compute so teams can triage regressions, explore physical design tradeoffs, and converge on signoff faster.

Ron Frenkel explains how the Azure Copilot Observability Agent (now generally available in Azure Monitor) helps teams investigate Azure AI Foundry and GenAI agent issues using Application Insights telemetry, including failures, latency, tool-call errors, token spikes, and dependency bottlenecks with evidence-backed root cause analysis.

Pamela Fox announces MCP Live, a free livestream focused on the Model Context Protocol (MCP), including updates from MCP maintainers and teams at Microsoft and GitHub, plus sessions on building MCP servers in VS Code and upcoming MCP authorization work.

Lee Stott invites AI engineers to a Microsoft Foundry Discord round table on using the Browser Automation Tool (BAT) to let agents drive real browser workflows via Playwright Workspaces, with a focus on setup basics, practical use cases, and the guardrails needed for responsible, auditable automation.

Jordan Selig explains how Microsoft Foundry’s new AI Gateway control plane lets platform teams create or associate an Azure API Management (APIM) gateway from the Foundry admin console, while keeping the runtime on Azure App Service. The post breaks down what Foundry now governs, what still belongs in APIM, and how to adapt an existing App Service agent sample.

VimalVerma outlines Hypervelocity Engineering (HVE) as an operating model for building and continuously evolving Azure AI Landing Zones, with a focus on platform engineering, Infrastructure as Code, Policy as Code, and security-by-design so enterprise AI platforms can scale without losing governance.

Lee Stott invites AI engineers to a Microsoft Foundry Discord round table on scaling agent apps beyond demos, focusing on how Foundry Toolbox, Skills, and Tool Search reduce tool sprawl, prompt bloat, and auth plumbing by centralizing tools behind a governed MCP endpoint with runtime discovery.

Rafia Aqil explains Microsoft’s IQ Platform (Work IQ, Fabric IQ, and Foundry IQ) and how it adds business and organizational context to AI systems. The post breaks down Fabric’s OneLake-based data layer, ontology-driven meaning, and Foundry IQ’s managed services for RAG, memory, ranking, and citations.

jordanselig explains the new stable Enterprise-Managed Authorization (EMA) extension for MCP and how it differs from a centrally governed OAuth setup using Microsoft Entra ID and Azure App Service Authentication. The post includes a working sample, a local EMA lab, and practical security details for deploying an Entra-governed MCP endpoint.

Noa Kuperberg breaks down how billing works for the Azure Copilot Observability Agent in Azure Monitor, including what gets metered, which operations are billable, where users can see per-response usage, and how to track costs in Azure Cost Management alongside standard Azure Monitor charges.

Manasa Ramalinga lays out a practical reference map for governing enterprise AI and autonomous agents, focusing on how to turn responsible AI policy into enforceable controls, runtime visibility, and audit-ready proof using Microsoft’s governance, security, and observability services.

jordanselig shares a reference implementation for giving an AI agent both short-term conversation history and durable, user-scoped memory on Azure App Service, using Redis for bounded session history and Cosmos DB vector search for recall, with keyless auth via managed identity and a one-command azd + Bicep deployment.

Krishna Roy shows how to load test Copilot Studio agents by simulating real multi-turn conversations over Direct Line (HTTP + WebSockets) with Locust, then running the same workload in Azure Load Testing. The post covers measuring TTFB vs full turn completion, handling turn.complete, file uploads, and secret handling with Key Vault.

Pamela Fox shows how to build an MCP server that returns more than plain text: image thumbnails as binary tool results and an interactive MCP app (a carousel) rendered inside VS Code, so GitHub Copilot can search, inspect, and present curated image results.

Lee Stott tours Microsoft’s open-source course on taking AI agents from prototype to production using Microsoft Agent Framework and Microsoft Foundry, covering agent design, multi-agent orchestration, evaluation, deployment, data sovereignty, and tool governance with concrete commands and code patterns.

jisunchoi explains how to replace “multi-model chaos” with a governed AI gateway on Azure using Azure API Management, covering cost controls (token quotas and budget-based model downgrades), security hardening (managed identity + private endpoints), observability with Application Insights, and a Terraform-based deployment you can integrate with GitHub Copilot.

abhimittal shows how to use Azure API Management (APIM) as an AI gateway in front of Azure AI Foundry to capture per-model token usage for governance. The post walks through an inbound policy that authenticates with managed identity, emits token metrics to Azure Monitor/Application Insights, and adds edge protection with Azure Front Door + WAF.

kinfey shares a reference implementation for running long-lived autonomous coding agents from Microsoft Teams, using MCP as the control plane and Azure Container Apps dynamic sessions as a Hyper-V-isolated sandbox. The post focuses on multi-agent orchestration, deployment reliability under platform timeouts, and practical security guardrails like auth, allowlists, private ingress, and managed identities.

supriyas lays out a practical, end-to-end lifecycle for building enterprise AI agents, using a banking “loan agent” example to show how to design guardrails, build with safety controls, test with evaluations and red teaming, deploy gradually, and continuously monitor and iterate using Microsoft Foundry and Azure services.

Arturo Quiroga shares major updates to the open-source Azure Architecture Diagram Builder, including a multi-turn Architecture Chat, a Blueprint (whiteboard-style) rendering mode, and a new MCP server interface so AI agents can generate, validate, cost, and render Azure architectures programmatically.

Nir Mashkowski shares customer examples of how Azure SRE Agent is being used to reduce incident triage and investigation time by having an AI-powered agent gather evidence, classify issues, and recommend next steps, with an emphasis on governance controls and operational “memory” for teams running production on Azure.