Uploaded June 2026 | Updated September 2026, 3 weeks ago
Descubre cómo implementar búsqueda semántica con vectores y aplicar el patrón Retrieval-Augmented Generation (RAG) para enriquecer las respuestas de los LLMs con información relevante y actualizada.
📍https://aka.ms/Microsoft.Extensions.AI
📌 Este evento es parte de una serie, aprende más aquí: https://aka.ms/MEAIseries
Capítulos:
0:02 Introducción y bienvenida
0:50 Introducción al concepto de Embeddings
2:14 El problema de las búsquedas tradicionales
3:21 ¿Qué son los Embeddings?
4:40 Relación semántica y distancia vectorial
6:26 Modelos de Embeddings (OpenAI)
7:47 Métodos para medir la similitud (Coseno, Euclidiana)
9:07 Casos de uso de la búsqueda vectorial
12:08 Demo: Creando el primer Embedding con Microsoft Extensions AI
20:49 Sesión de preguntas y respuestas (IDE y herramientas)
25:15 Almacenamiento de vectores y base de datos
29:52 Implementación de clase Product y Vector Store
33:17 Demo: Búsqueda vectorial práctica
37:55 Ejemplo de búsqueda con Tensor Primitives
42:43 Integración con SQL Server y búsqueda vectorial
50:41 Implementación del patrón RAG en Azure Foundry
52:21 Plantilla de chatbot con RAG (Blazor)
54:11 Demo del chatbot con RAG sobre documentos
56:11 Conclusión y recursos adicionales
#microsoftreactor #learnconnectbuild
[EventID:27062]
Descubre cómo implementar búsqueda semántica con vectores y aplicar el patrón Retrieval-Augmented Generation (RAG) para enriquecer las respuestas de los LLMs con información relevante y actualizada.
📍https://aka.ms/Microsoft.Extensions.AI
📌 Este evento es parte de una serie, aprende más aquí: https://aka.ms/MEAIseries
Capítulos:
0:02 Introducción y bienvenida
0:50 Introducción al concepto de Embeddings
2:14 El problema de las búsquedas tradicionales
3:21 ¿Qué son los Embeddings?
4:40 Relación semántica y distancia vectorial
6:26 Modelos de Embeddings (OpenAI)
7:47 Métodos para medir la similitud (Coseno, Euclidiana)
9:07 Casos de uso de la búsqueda vectorial
12:08 Demo: Creando el primer Embedding con Microsoft Extensions AI
20:49 Sesión de preguntas y respuestas (IDE y herramientas)
25:15 Almacenamiento de vectores y base de datos
29:52 Implementación de clase Product y Vector Store
33:17 Demo: Búsqueda vectorial práctica
37:55 Ejemplo de búsqueda con Tensor Primitives
42:43 Integración con SQL Server y búsqueda vectorial
50:41 Implementación del patrón RAG en Azure Foundry
52:21 Plantilla de chatbot con RAG (Blazor)
54:11 Demo del chatbot con RAG sobre documentos
56:11 Conclusión y recursos adicionales
#microsoftreactor #learnconnectbuild
[EventID:27062]
![Code with AI: GitHub Copilot로 시작하는 AI-Native 개발
세션 소개
AI는 소프트웨어 개발 방식을 빠르게 바꾸고 있습니다. 개발자는 AI와 함께 협업하며 코드를 작성하고, 반복 작업을 줄이며, 개발 속도를 높이고 있습니다.
이번 세션에서는 GitHub Copilot이 지원하는 새로운 AI-Native 개발 워크플로우를 실시간 데모와 함께 살펴봅니다.
이슈 관리부터 코드 작성, 변경 사항 검토, Pull Request 생성까지 AI와 함께 개발하는 과정을 직접 확인하며, 아이디어를 실제 코드로 구현하는 과정을 더욱 빠르고 효율적으로 만드는 방법을 소개합니다.
세션에서는 이런 내용을 다룹니다
- AI-Native 개발 방식이 가져온 소프트웨어 개발의 변화
- GitHub Copilot을 활용한 이슈 관리, 코드 작성, Pull Request 워크플로우
- AI를 활용해 코드 작성과 반복 개발을 더욱 빠르게 수행하는 방법
- 개발 전 과정에서 AI와 효과적으로 협업한 사례
주요 기술
- GitHub Copilot App
- GitHub Copilot (Chat & CLI)
- AI-Native 개발 워크플로우
- Cloud 및 Local Sandbox
- AI 기반 코드 생성 및 반복 개발
추천 대상
- AI-Native 개발 방식을 도입하고 싶은 개발자
- GitHub Copilot을 개발 업무에 활용하고 있는 엔지니어
- AI를 활용해 개발 생산성을 높이고 싶은 소프트웨어 개발자
- 최신 AI 기반 개발 경험에 관심 있는 모든 개발자
[eventID:27479] Code with AI: GitHub Copilot로 시작하는 AI-Native 개발](https://i.ytimg.com/vi/Ww9Io4dcu-s/mqdefault.jpg)
![Estruturando Agentes com AGENTS.md e Prompt Files
Estruturação do comportamento de agentes utilizando AGENTS.md e prompt files reutilizáveis. Organização de instruções por tarefa, definição de responsabilidades e evolução do setup para um modelo modular e reutilizável.
📌 Este evento faz parte de uma série, saiba mais aqui: https://aka.ms/AgenticDevVSCode
🔗 https://aka.ms/AgenticDevSeries/Resource
[eventID:27388] Estruturando Agentes com AGENTS.md e Prompt Files](https://i.ytimg.com/vi/XZnLvDWCx1Y/mqdefault.jpg)
![Ship with AI: Review, Secure, and Deploy with Confidence
Building software is only part of the journey. Delivering secure, reliable applications at scale requires confidence across the entire development lifecycle.
In this session, youll learn how to bring AI into real team workflows, from code review and security validation to CI/CD and automated merging. Through live demonstrations, youll see how AI helps improve quality, strengthen security, and maintain consistency across every stage of development, enabling teams to confidently ship AI-powered and agent-driven systems.
You Will Learn
• How AI enhances code review and development quality
• How to incorporate AI into security and compliance workflows
• How to automate deployment processes using modern CI/CD practices
• How to confidently ship AI-powered applications at scale
Technologies Used
• GitHub Copilot Code Review
• Agent Merge
• GitHub Actions (CI/CD)
• AI-assisted security review
• End-to-end development lifecycle automation
Who Should Attend
• Developers responsible for code quality and deployment
• DevOps and platform engineers managing CI/CD workflows
• Engineering teams building AI-powered applications
• Anyone interested in using AI to review, secure, and ship software with confidence
[eventID:27465] Ship with AI: Review, Secure, and Deploy with Confidence](https://i.ytimg.com/vi/XaoojoTQHJc/mqdefault.jpg)
![How Microsoft 1ES uses agentic AI to take on security and compliance at scale
Demo heavy content on how we are building agentic workflows:
How Microsoft 1ES uses agentic AI to take on security and compliance at scale.
How we build and use Azure SRE Agent with agentic workflows From Copilots to Coworkers:
How AI Agents Are Transforming Azure Networking Operations
How Microsoft 365 built a platform engineering layer on AKS to ship faster at global scale Running Foundry Agent Service on Azure Container Apps.
[eventID:27445] How Microsoft 1ES uses agentic AI to take on security and compliance at scale](https://i.ytimg.com/vi/Xqci_OtmeCU/mqdefault.jpg)
![Financial Audit Automation Using Microsoft Copilot Studio & Code Interpreter
In this session, participants will learn how to leverage Microsoft Copilot Studio and Code Interpreter to automate financial audit processes and gain actionable insights from financial data. The session focuses on transforming traditional manual audit activities into intelligent, AI-driven workflows that improve accuracy, efficiency, and decision-making.
Participants will discover how to build an AI-powered audit assistant capable of analyzing financial statements, identifying anomalies, detecting inconsistencies, generating audit summaries, and answering natural language questions about financial records.
Key Topics Covered:
- Introduction to AI-powered financial auditing
- Understanding Microsoft Copilot Studio for business automation
- Using Code Interpreter for financial data analysis
- Uploading and analyzing financial reports and spreadsheets
- Detecting anomalies, outliers, and compliance risks
- Generating automated audit findings and summaries
- Creating conversational audit assistants
- Building end-to-end audit workflows with AI Hands-On Project: Build a Financial Audit Copilot that can review financial data, identify potential issues, generate audit reports, and provide insights through a conversational interface.
Learning Outcomes:
- Understand AI applications in financial auditing
- Analyze financial data using Code Interpreter
- Create intelligent audit assistants with Copilot Studio
- Automate audit reporting and compliance checks
- Reduce manual effort while improving audit accuracy By the end of this session, participants will have a working AI-powered financial audit solution and a clear understanding of how to apply Copilot Studio and Code Interpreter to real-world finance and compliance scenarios.
[eventID:27358] Financial Audit Automation Using Microsoft Copilot Studio & Code Interpreter](https://i.ytimg.com/vi/XvJnY1clKa4/mqdefault.jpg)
![¿Y si tu próximo agente fuera una Logic App? Conversacional, autónomo y MCP Server
Cuando pensamos en construir agentes de IA, pensamos en frameworks, SDKs y todo el boilerplate que hay que montar antes de que el agente haga algo útil. Pero hay otro camino: Logic Apps ha incorporado el agent loop a su motor de workflows, ahora todo lo que ya tenía (conectores, ejecución durable, retries, monitorización, managed identities) se convierte en infraestructura de agentes que te viene dada.
En esta sesión, creamos en directo sobre un mismo escenario, un asistente de soporte interno, para poder compararlas bien: Parte 1, el agente conversacional. Creamos el agente desde el designer, le conectamos el modelo y le damos sus primeras tools usando conectores. Vemos cómo se depura de verdad: run history, transcripción del chat, y qué entra y sale del modelo en cada vuelta del loop. Y resolvemos el problema que casi nadie cuenta: el contexto de usuario.
Con autorización on-behalf-of, las tools dejan de ejecutarse con una cuenta de servicio y pasan a actuar con la identidad del usuario que está chateando: si el agente manda un correo, sale de tu buzón. Desplegamos el agente en Microsoft Teams. Parte 2, el agente autónomo.
Mismo escenario, pero sin nadie al otro lado: un trigger de evento arranca el workflow, el agente clasifica, decide y ejecuta por su cuenta. Vemos dónde colocar las agent actions dentro de un workflow normal, cómo combinar varias, y qué pasa con las ejecuciones de larga duración, los errores y la gobernanza cuando no hay un humano mirando.
Parte 3, la Logic App como MCP Server. Le damos la vuelta a la tortilla: en lugar de que el agente viva en Logic Apps, exponemos nuestros workflows como tools MCP para que los consuma cualquier agente externo, desde VS Code hasta el agente que tú quieras. Qué requisitos tiene que cumplir un workflow para ser una tool, cómo se gestionan las API keys y cómo montar varios MCP Servers en una misma Logic App.
Además, hablaremos sobre decisiones de arquitectura: Consumption con modelo gestionado por Microsoft o Standard trayendo tu propio modelo de Foundry, con sus diferencias en redes, autenticación, billing y desarrollo local. Si vienes del mundo pro-code, esta sesión te va a sorprender: no se trata de elegir entre código o designer, sino de saber cuándo el motor de Logic Apps te ahorra construir la mitad de tu plataforma de agentes.
Key Takeaways: Cómo construir agentes conversacionales y autónomos con el agent loop de Logic Apps, y en qué se diferencian Cómo ejecutar tools con la identidad real del usuario mediante autorización on-behalf-of Cómo desplegar tu agente en Microsoft Teams para que hable donde ya están tus usuarios Cómo exponer workflows como MCP Servers remotos consumibles desde cualquier agente o cliente MCP Consumption o Standard: qué SKU elegir según modelo, redes, billing y compliance.
🔗 Aprende más accediendo a los recursos: https://aka.ms/AgentsLogicApps
#MicrosoftReactor #LearnConnectuild #LogicApps #MCP
[eventID:27563] ¿Y si tu próximo agente fuera una Logic App? Conversacional, autónomo y MCP Server](https://i.ytimg.com/vi/Y-TNojksass/mqdefault.jpg)

![Architecting Context-Aware Agents with the Microsoft IQ Stack
Context-aware agents depend on a design that connects knowledge, enterprise data, work signals, and the live web.
Learn how the Microsoft IQ stack (Foundry IQ, Fabric IQ, Work IQ, and Web IQ) provides the enterprise intelligence layer for AI agents.
Map each layer onto a reference architecture, see where each one fits and how they compose, and learn how to choose the right intelligence for a given scenario.
Build a practical foundation for designing scalable, reliable, context-aware agents.
00:00 Welcome & Introduction
01:14 Why Context Matters for AI Agents
02:50 Understanding the Microsoft IQ Stack
04:07 Web IQ Overview
05:08 Web IQ Demo: Real-Time Grounding
07:57 Foundry IQ and Agentic Knowledge Retrieval
09:42 Fabric IQ and Business Data Understanding
11:13 Building a Hotel Agent with Foundry, Fabric, and Work IQ
17:27 Semantic Models vs Ontologies
22:54 Fabric Data Agents and Multi-Source Retrieval
26:30 Work IQ and Microsoft 365 Data
29:19 Agent 365 Autopilots Explained
33:29 Demo: Work IQ in Microsoft 365, GitHub Copilot, and Scout
37:30 Agent Identities, Permissions, and Security
44:31 Q&A: Permissions, Licensing, and Deployment Gotchas
52:34 Microsoft IQ Learning Resources and Upcoming Sessions
59:45 Closing Remarks
🔗 Resources: https://aka.ms/iq-series
📌 This event is a part of a series, learn more here: https://aka.ms/microsoftiqlive
[eventID:27386] Architecting Context-Aware Agents with the Microsoft IQ Stack](https://i.ytimg.com/vi/YOHnkoQ4Jic/mqdefault.jpg)

![Get Certified: Which Data Exam Fits You Best?
Not sure where to start? With options like PL-300, DP-600, DP-700, DP-800, and DP-900, it is easy to feel stuck before you even begin.
In this session, we will break down each exam and the role it maps to, from Power BI Data Analyst to Fabric Analytics Engineer, Fabric Data Engineer, and SQL AI Developer. You will get a clear sense of what each exam covers, how they differ, and which one fits your background, interests, and goals.
We will also share practical tips on how to choose your path with confidence so you can avoid spending time on the wrong thing and focus on what actually moves you forward. By the end, you will know exactly where to start and what to do next.
🔗 Register for Data Days: https://aka.ms/datadays
📍 This session is a part of a series. Learn more here: https://aka.ms/datadayslive/cert-y
#MSFTReactor #learnconnectbuild #DataDays #GetCertified
Chapter markers:
0:00 Introduction and housekeeping
0:58 Speakers introduction
2:00 Data Days overview and resources
3:10 Fabric community and contests
4:06 Power BI Data Viz World Championship
4:53 User panels
6:34 Moderator introductions
7:12 Agenda overview
8:14 Microsoft certifications explained
9:45 DP-900: Azure Data Fundamentals
11:30 Role-based certifications overview
12:46 PL-300: Power BI Data Analyst
15:07 DP-600: Fabric Analytics Engineer
19:07 DP-700: Fabric Data Engineer
23:46 DP-800: SQL AI Developer Associate
29:05 Q&A: Experience requirements
30:28 Q&A: Overlap between exams
32:28 Q&A: Job roles for DP-800
33:36 Q&A: Exam resources and strategy
35:10 Deciding which exam is right for you
36:18 Additional sessions and closing resources
37:54 Q&A: Deep dive into exam content and recommendations
42:33 Q&A: Study groups
44:24 Q&A: Why get certified?
47:11 Q&A: Exam format and structure
48:36 Q&A: Exam order recommendations
50:11 Closing remarks
[eventID:27217] Get Certified: Which Data Exam Fits You Best?](https://i.ytimg.com/vi/YViNorNYV3c/mqdefault.jpg)
![Give Microsoft AI Agents Communication Superpowers
In this webinar, youll learn how to use Infobips MCP servers to give your AI agents running on Azure communication tools.
From personal Al assistants to scalable customer support and marketing agents, discover how to build agents that reach, engage, and deliver outcomes through global communications.
🔗 https://aka.ms/UnlockAIAgentCommunication/Blog
🔗 https://aka.ms/MCPcenter
🔗 https://aka.ms/AIagentsAzure
Explore the documentation for the three showcased examples below:
🔗 Weather Agent: https://www.infobip.com/docs/tutorials/build-sms-weather-agent-copilot-studio
🔗 GitHub Issue Agent: https://www.infobip.com/docs/tutorials/build-github-security-alert-agent
🔗 Trivia Agent: https://www.infobip.com/docs/tutorials/build-whatsapp-trivia-agent
Ready to give your Microsoft AI agents communication superpowers? 🚀
0:13 - Introduction to the Session
1:03 - What is Infobip & Communication Platforms
3:52 - Understanding MCP (Model Context Protocol)
6:44 - Overview of the Three AI Agent Demos
7:31 - Demo 1: Weather Agent (using Copilot Studio)
17:08 - Demo 2: GitHub Security Issue Agent (using Microsoft Foundry)
32:12 - Demo 3: WhatsApp Trivia Agent (using Microsoft Foundry)
54:10 - Resources, Summary, and Q&A
#MicrosoftReactor #LearnConnectBuild
[eventID:27362] Give Microsoft AI Agents Communication Superpowers](https://i.ytimg.com/vi/Yht3hOf-MQU/mqdefault.jpg)