YouTube Microsoft Developer

AI agent costs grow faster than the work itself, and caching is how you bring them back down. Gareth Bland, Chief Data Scientist at Microsoft, explains. From our full conversation on TokenOps (https://youtu.be/B4yovi_znxE). #AIAgents #TokenOps

8. 10. 2026

https://aka.ms/foundry-portal https://aka.ms/InsideMicrosoftFoundryPlaylist Your agent took too long to answer, or it answered with confidence and got it wrong. Was it the model, a tool, or the instructions? Without traces, you're guessing. With traces, you connect Application Insights to your Microsoft Foundry project, which turns on server-side tracing with no code changes. Then you read an agent run span by span: the agent invocation, each model call, and every tool call with its arguments, result, tokens, and latency. You find the slow step and the wrong step, then query the same data in Application Insights with KQL to see if it's one bad run or a pattern. By the end, you'll know exactly where to look first. 0:00 - Connect Agents to Enterprise Context 1:09 - Explore Web IQ 1:45 - Ground Responses with Foundry IQ 2:51 - Connect Data with Fabric IQ 3:57 - Query Semantic Models and Ontologies 4:33 - Take Actions with Work IQ 5:38 - Give the Agent Its Own Identity 6:11 - Automate Supply Escalations 7:18 - Manage Agents with Agent 365 8:22 - Add Teams Messaging Support

8. 10. 2026

Learn all about what's new after the conference! ✅ Chapters: 0:00 Intro and background on SQLCon/FabCon in Barcelona 2026 1:04 Hybrid and migration updates 3:12 Database Hub updates 3:40 Hyperscale updates 4:30 SQL DB in Fabric updates 6:00 Tools, agents, developer updates 8:15 Across the versionless engine updates ✅ Resources: https://aka.ms/sqlconnews https://aka.ms/mssqldecks https://aka.ms/sqlconeudemos https://aka.ms/sqlroadmap ✅ Let's connect: Twitter - Anna Hoffman, https://twitter.com/AnalyticAnna Twitter - AzureSQL, https://aka.ms/azuresqltw 🔴 Watch even more Data Exposed episodes: https://aka.ms/dataexposedyt 🔔 Subscribe to our channels for even more SQL tips: Microsoft SQL / Azure SQL: https://aka.ms/msazuresqlyt Microsoft SQL Server: https://aka.ms/mssqlserveryt Microsoft Developer: https://aka.ms/microsoftdeveloperyt #AzureSQL #SQL #LearnSQL

8. 10. 2026

AI agent costs don't grow in step with the work. As an agent takes more turns, context piles up and uncached input costs rise on a quadratic curve: in Gareth Bland's example, an agent that takes four times as many turns costs sixteen times as much. Engineering your context to use cached input can bring that curve down to a tenth of the cost. Gareth Bland, Chief Data Scientist at Microsoft, explains TokenOps: the practice of deciding where token spend will move the needle most. He covers why reasoning makes output tokens the expensive part, how to measure agentic efficiency with three ratios (autonomous completion rate, token efficiency ratio, and value generation ratio), and how an S-curve shows whether your team is underinvesting, getting real leverage, or gold-plating features nobody asked for. Resources - Unit Economics of Agentic AI https://aka.ms/tokenomicon-2026 - Maximize ROI from AI https://azure.microsoft.com/solutions/maximize-roi-from-ai - More Azure resources! https://azure.com/AzureEssentials Connect - Gareth Bland https://www.linkedin.com/in/gareth-bland/ Chapters 0:00 What TokenOps means for AI agent spending 0:35 Why output tokens cost more than input tokens 1:02 Why AI agent costs grow faster than the work 1:50 How caching brings AI agent costs down 2:05 How to measure agentic efficiency 3:05 How to know where your next token dollar belongs

7. 10. 2026

Welcome to the next MVP Unplugged, where Microsoft MVPs share real-world projects and insights from the field! In this episode, host Justin Garrett sits down with Microsoft MVP Mahmoud A. Atallah to explore an Azure Delivery Factory custom agent that turns meeting transcripts, proposals, PDFs, and briefs into structured requirements and execution-ready project artifacts. Mahmoud walks through his process in GitHub Copilot and VS Code: convert source files into structured Markdown; extract scope and requirements; use custom agents, skills, and scripts to generate project plans, runbooks, task lists, risks, kickoff questions, executive summaries, and reusable delivery artifacts; then enrich the results with Azure MCP Server and Microsoft Learn MCP Server guidance aligned to the Well-Architected Framework. He also explains how reusable JSON, Python, and PowerShell automation, thoughtful model selection, and human validation improve speed, consistency, and trust. ⭐ What You’ll Learn: ✅How to transform raw meeting transcripts, proposals, PDFs, and briefs into structured project requirements ✅How GitHub Copilot custom agents, skills, and scripts create repeatable delivery workflows ✅How to generate project plans, runbooks, task lists, risks, kickoff questions, and executive summaries ✅How Azure MCP Server and Microsoft Learn MCP Server add current Azure guidance and Well-Architected Framework alignment ✅Why reusable JSON, Python, and PowerShell automation can reduce LLM token usage and processing time ✅How model selection, command approvals, and human validation keep AI-assisted delivery under control ✅How repository context and existing project artifacts help an agent update work instead of starting from scratch ✅How multi-agent orchestration could extend delivery automation into community, events, marketing, and communications ✅ Chapter Markers 00:00 – Intro to MVP Unplugged 00:20 – Meet Microsoft MVP Mahmoud A. Atallah 01:13 – Azure Delivery Factory: From Meeting Transcripts to Project Plans 04:16 – GitHub Copilot Custom Agents, Skills, and MCP Servers 08:47 – Turning Customer Calls into Structured Requirements 13:20 – Demo: Discovery Transcript to Delivery Artifacts 17:36 – Demo: Technical Proposal PDF to Project Plan 20:19 – JSON, Python, and PowerShell Automation for Faster Agents 23:57 – Reviewing Project Plans, Tasks, Risks, and Owners 27:47 – Multi-Agent Orchestration and Mahmoud’s MVP Journey 👥 Speakers Mahmoud A. Atallah is a Microsoft MVP and Azure Solution Leader at Bespin Global. Originally from Egypt and based in Dubai, Mahmoud specializes in Azure AI and Azure Virtual Desktop. He builds practical AI-assisted delivery workflows that use GitHub Copilot custom agents, skills, scripts, MCP servers, and reusable automation to turn unstructured customer inputs into consistent, execution-ready project artifacts. Follow Mahmoud on LinkedIn: https://www.linkedin.com/in/mahmoudatallah Justin Garrett is host of MVP Unplugged, Principal PM in Developer Relations which is part of Microsoft Cloud + AI. Justins career at Microsoft also spans 20 years across Windows, Bing, Edge, Web Platform, Students/ University Relations, Cloud Advocacy, and most recently a leader of the MVP Program at Microsoft. Follow Justin on LinkedIn: https://www.linkedin.com/in/justgar/. About MVP Unplugged AI is reshaping how we work and live. And for developers and technologists alike, the pace of innovation, new tools, new models, patterns & practices, and even culture itself is changing even faster. It can be difficult to know what to learn, what to prioritize, what truly lives up to the promise of unlocking creativity and boosting productivity. Join Justin Garrett, Principal PM in DevRel and leader in the Microsoft MVP Program as he speaks with MVPs to share what they're learning using a real-world project in this conversational series. In each episode, they'll experiment, code, and share honest insights that can make a real difference for the audience. Justin and his guests share stories of navigating technological change and look ahead for what's next in tech. Come discover with us how to thrive in this era of AI! 🔗 Resources & Links 💻 Mahmoud's AI Agent Skills repository https://github.com/3tallah/awesome-ai-agent-skills 🏗️ Azure Well-Architected Framework https://learn.microsoft.com/azure/well-architected 🔌 Azure MCP Server https://learn.microsoft.com/azure/developer/azure-mcp-server/ 📚 Microsoft Learn MCP Server https://learn.microsoft.com/training/support/mcp 🚀 Try GitHub Copilot https://github.com/features/copilot 🔔 Subscribe for more MVP stories, AI engineering walkthroughs, and handson developer content! #MicrosoftDeveloper #MVPUnplugged #GitHubCopilot #CustomAgents #AzureAI #MCPServer #MicrosoftLearn #WellArchitectedFramework #ProjectManagement #AIAutomation #MicrosoftMVP

6. 10. 2026

https://aka.ms/foundry-portal https://aka.ms/InsideMicrosoftFoundryPlaylist A smart model without the right context still gives the wrong answer. See how Microsoft IQ connects your agent to business data, retrieves relevant context at runtime, and grounds each response in the information that matters. 0:00 - Connect Agents to Enterprise Context 1:09 - Explore Web IQ 1:45 - Ground Responses with Foundry IQ 2:51 - Connect Data with Fabric IQ 3:57 - Query Semantic Models and Ontologies 4:33 - Take Actions with Work IQ 5:38 - Give the Agent Its Own Identity 6:11 - Automate Supply Escalations 7:18 - Manage Agents with Agent 365 8:22 - Add Teams Messaging Support

6. 10. 2026

In this episode of the Azure Essentials Show, Maria Jose and Kyle Ikeda look at a common challenge for FinOps and architecture teams: controlling database costs while the environment keeps changing. The conversation introduces the savings plan for databases, a flexible, spend-based way to get discounts that follow your workloads across services and regions. They also cover how it fits next to Azure Reservations, so teams can match each workload with the right cost strategy, and finish with a demo and practical steps for getting started. In this episode, you will learn… - How the savings plan for databases gives you flexible discounts that follow your workloads across services and regions - When to use a savings plan, when to use Azure Reservations, and how they work together - How to get started, from getting a recommendation that fits your usage to buying and tracking your plan Resources - Azure Advisor https://azure.microsoft.com/products/advisor - Savings plans https://azure.microsoft.com/pricing/offers/savings-plans - More Essential resources! https://azure.com/AzureEssentials - Azure Portal https://portal.azure.com Related episodes - Azure Cost Estimation: Navigate Database Pricing https://aka.ms/AzEssentials/247 - Azure Cost Estimation: Plan Confidently with the Azure Pricing Calculator https://aka.ms/AzEssentials/231 - Watch more Azure Pricing episodes https://aka.ms/AzurePricingVideos - Watch the Azure Essentials Show https://aka.ms/AzureEssentialsShow Connect - Maria Jose Fernandez https://www.linkedin.com/in/mariajosefm/ - Kyle Ikeda https://www.linkedin.com/in/kyle-ikeda-49ab7a30/ Chapters 00:00 In this episode … 00:17 Introduction 00:54 Why is cost management harder? 01:37 What is the savings plan for databases? 01:59 How does the savings plan apply to a changing environment 02:38 How does it work 03:25 Comparison to Azure reservations 04:27 Purchase savings plans in Azure Advisor 05:23 Purchase savings plans in the Azure portal 06:31 How to get started 07:08 Ask us anything

6. 10. 2026

🚀 Save your spot: https://aka.ms/CopilotStudio/Skilling 📅 October 8, 2026 🕒 8:00 AM - 9:00 AM (UTC-07:00) Pacific Time (US & Canada) Discover what to learn first, how to get started building with Copilot Studio, and how Agent Academy can help you develop practical AI builder skills. #CopilotStudio #AI #Agents #AgentAcademy

4. 10. 2026

We're ending Universe the way it started: with the people who build in the open. Join us for a final session of open source voices, unexpected demos, and the kind of nerdy joy that reminds you why you got into this in the first place. More details coming soon. #GitHub #GitHubUniverse Stay up-to-date on all things GitHub by connecting with us: YouTube: https://gh.io/subgithub Blog: https://github.blog X: https://twitter.com/github LinkedIn: https://linkedin.com/company/github Insider newsletter: https://resources.github.com/newsletter/ Instagram: https://www.instagram.com/github TikTok: https://www.tiktok.com/@github About GitHub It’s where over 180 million developers create, share, and ship the best code possible. It’s a place for anyone, from anywhere, to build anything—it’s where the world builds software. https://github.com

2. 10. 2026

Join us for the opening keynote as we look at the next era of software, the developers building it, and the tools and workflows GitHub is creating to help them go further. Expect new announcements, live demos, and a few surprises along the way. #GitHubUniverse #GitHubCopilot #GitHub Stay up-to-date on all things GitHub by connecting with us: YouTube: https://gh.io/subgithub Blog: https://github.blog X: https://twitter.com/github LinkedIn: https://linkedin.com/company/github Insider newsletter: https://resources.github.com/newsletter/ Instagram: https://www.instagram.com/github TikTok: https://www.tiktok.com/@github About GitHub It’s where over 180 million developers create, share, and ship the best code possible. It’s a place for anyone, from anywhere, to build anything—it’s where the world builds software. https://github.com

2. 10. 2026

https://aka.ms/foundry-portal https://aka.ms/InsideMicrosoftFoundryPlaylist Model Router routes to the best model for the task. But, before you switch, you'll want to know: will it save me money on my workload without hurting quality? In this episode we use the open-source Model Router Auto Evaluation toolkit. It compares Model Router with your current baseline model (for example GPT-5) on: quality, cost, latency, value and model distribution. At the end you get a HTML dahsboard with 8 charts which answers the question: do I need a giant model or route a selection of the models using the Model Router? 0:00 - Save Costs with Model Router 0:35 - Model Router Auto Evaluation 1:08 - Configure Models and Pricing 1:39 - Prepare the Evaluation Dataset 2:10 - Run and View the Evaluation 2:40 - Compare Cost, Latency, and Quality 3:12 - Analyze the Quality Breakdown 3:43 - Choose the Right Tradeoffs

1. 10. 2026

Learn about vector indexing in Microsoft SQL from the PM and engineering team. ✅ Chapters: 0:00 Intro to Product Management and Engineering vector team 1:13 What things were customers asking for. Example scenario: support team looking for similar or related cases 2:02 Everything's fine but now it needs to scale better, so we built and index 2:37 Anna asks why can't we just use exact search? 2:59 Intro to DiskANN - a graph-based vector index 3:35 Why is it called DiskANN 4:30 Next customer asks for not just vector search but also iterative filtering 5:50 Anna reacts to iterative filtering vs post-filtering 6:33 Next customer asks for insert, update, delete, without needing to update the index 7:05 DML vector demo 8:00 Approximate vs exact vector search, handled by optimizer 10:20 Next customer asks to search on 1 billion rows 10:49 Vector search still returns within milliseconds 11:17 Pooja reacts to all the customers asks being met so far 12:05 Vector Index in Azure SQL PaaS and Fabric SQL is now generally available! ✅ Let's connect: LinkedIn - Pooja Kamath LinkedIn - Krithika Subramanian Twitter - Anna Hoffman, https://twitter.com/AnalyticAnna Twitter - AzureSQL, https://aka.ms/azuresqltw 🔴 Watch even more Data Exposed episodes: https://aka.ms/dataexposedyt 🔔 Subscribe to our channels for even more SQL tips: Microsoft Azure SQL: https://aka.ms/msazuresqlyt Microsoft SQL Server: https://aka.ms/mssqlserveryt Microsoft Developer: https://aka.ms/microsoftdeveloperyt #AzureSQL #SQL #LearnSQL

1. 10. 2026

Discover the new developer experience for extending Microsoft 365 Copilot with Copilot Plugins and Work IQ Developer Tools (WIQD). See how agents, skills, connectors, and MCP capabilities can come together as plugins, and how WIQD provides an agentic, end-to-end experience for building, validating, publishing, and monitoring them.

30. 9. 2026

What happens when Copilot goes beyond answering questions and starts doing the work with you and for you? Discover the latest Copilot innovations, including Cowork, Autopilot (formerly Scout), and more, and see how Microsoft is moving towards a more proactive, agentic way of working. Introducing the future of Copilot.

30. 9. 2026

Microsoft 365 Copilot is becoming much more than a chat experience. Copilot Chat, Copilot Cowork, Work IQ, Autopilots, Agent and Skill Builder, SharePoint Agents, Scout, GitHub Copilot, Agent 365, and more are forming an ecosystem where people, agents, organisational knowledge, and tools can work together. Join us for a tour of this evolving landscape as we connect the dots between these technologies and explore the journey from asking AI for help to delegating work and, ultimately, enabling agents that can proactively keep work moving.

30. 9. 2026