Browse Artificial Intelligence Community (163)

skundapura explains how to observe and troubleshoot AI agent workloads in Azure AI Foundry, comparing tracing and telemetry support across Microsoft Agent Framework, Semantic Kernel, LangChain/LangGraph, and the OpenAI Agent SDK using OpenTelemetry with Azure Monitor and Application Insights.

dbandaru shows how to connect the New Relic-hosted MCP server (Preview) to Azure SRE Agent, including API key setup, RBAC considerations, and how to create a New Relic-focused subagent/skill so the agent can query NRQL, traces, logs, metrics, alerts, and dashboards via natural language.

VaidhyaP introduces AG-UI (Agent–User Interface), a protocol for connecting AI agents to rich frontends with streaming events, declarative UI proposals, shared state updates, and human-in-the-loop approvals, plus practical security guidance like Azure AD protection and Key Vault-backed secrets.

Samarpitaa explains where Azure AI Foundry IQ fits (and doesn’t) for enterprise agent knowledge access, then shows a reference approach for querying Foundry IQ knowledge bases directly via the Azure AI Search Python SDK with permission-aware retrieval and citations.

Shah_Viral explains how to build an enterprise “knowledge copilot” on Azure using Foundry IQ knowledge bases and Azure AI Search agentic retrieval, including C#/.NET setup, MCP-based agent connection, and key trade-offs around preview maturity, cost, latency, and security controls like ACLs and Purview labels.

dikshashakya explains how to turn long-form video transcripts into structured Standard Operating Procedures (SOPs) using GraphRAG to build a knowledge graph and Azure OpenAI to generate grounded sections like scope, definitions, responsibilities, and step-by-step procedures.

JennyF explains how Microsoft’s 1ES team uses agentic AI (including GitHub Copilot CLI) plus “skills” and “agent signals” to speed up CVE remediation and compliance work across many repositories, while keeping humans in the loop for review, validation, and deployment.

vsriramdas explains how to use Microsoft PyRIT to red-team agentic AI systems, then shows how to wrap PyRIT with a YAML-driven CLI so you can run repeatable scans in CI/CD and gate releases based on OWASP LLM Top 10-aligned findings.

lexinadolski recaps Microsoft’s presence in the CNCF Project Pavilion at KubeCon EU 2026, summarizing the technical conversations and themes across Kubernetes projects—migration to Gateway API, confidential computing, image signing, observability tooling, and requests for deeper Azure/AKS and AI-workload support.

rajesh-yadav breaks down what shipped in Microsoft Agent Framework 1.0 (GA), explaining its agent + workflow split, core runtime building blocks, and new interoperability pieces like A2A and MCP. The post includes minimal C# and Python examples using Azure AI Foundry/Projects endpoints to run an agent in production-style setups.

Sreekanth_Thirthala announces AI skill assessment in Azure API Center, a built-in quality scoring feature using an LLM-as-a-judge approach. It adds automated pass/fail scoring, per-dimension quality metrics, and structural/schema checks so teams can govern and adopt production-ready skills with more confidence.

B_Manasa explains how GitHub Copilot (especially Copilot Chat in VS Code) can speed up relational data modeling by turning architecture intent into reviewable schema drafts faster, using a multi-tenant SaaS control-plane example and concrete prompt patterns for iterating on cardinality, history tables, and schema evolution.

Parvathy_R_Pillai compares traditional ML pipelines with Azure AI Foundry, focusing on the shift from model-centric delivery to operating end-to-end AI applications (including agents) with built-in governance, evaluation, and observability for production use.

kmalkov shares a real-world fintech lending ML decisioning workload evaluated using Microsoft’s Analog Optical Computer (AOC) digital twin on Azure, focusing on production-scale volumes, weighted ensemble models, and end-to-end explainability and auditability for credit, affordability, and risk decisions.

kumar_rahul introduces Microsoft’s in-house MAI models for speech-to-text, text-to-speech, and text-to-image, and explains what changes for Azure developers—especially around Foundry-native governance (RBAC, Entra ID, Managed Identity) and building agent-oriented, multimodal workloads.

PeterTHLee shares a validated Azure reference architecture for drone-based industrial inspections that combines deterministic computer vision with Azure OpenAI reasoning. The post breaks down an event-driven pipeline (Blob Storage → Functions → Vision/AML → OpenAI → Foundry evaluation → Cosmos DB → Power BI) and calls out security controls needed for production use.

Jon_Andoni_Baranda explains how the Azure Compute team uses AI and Model Context Protocol (MCP) to automate downtime investigations for Azure VMs by running Kusto queries across telemetry sources, building a recovery timeline, and attaching a structured root-cause report to incident tickets in minutes.

kirankumar_manchiwar04 explains how to run Azure OpenAI in a Zero Trust setup by removing public exposure and routing traffic privately through Azure VNets, Private Endpoints, and Private DNS. The post includes an end-to-end reference architecture and a step-by-step configuration checklist to validate private connectivity.

ankitasarkar explains why a pure RAG approach can produce inconsistent or logically wrong matches in enterprise document mapping, and how adding a knowledge-graph layer to constrain retrieval improves consistency, relevance, and explainability.

kinfey compares Anthropic’s “Managed Agents” architecture with Microsoft Foundry Hosted Agents, then walks through a Python sample that combines a stateless orchestrator, replaceable execution sandboxes, and an append-only event log, with Azure-native hosting, observability, and identity.