Microsoft Launches Agent Framework (Preview) to Simplify AI Agent Development

Microsoft has introduced the Agent Framework (Preview), giving developers a new way to build and run AI agents within .NET. The framework is designed to reduce complexity, making it easier to connect agents, manage orchestration, and streamline hosting. This marks a step toward making AI agent development more accessible to every developer.

Here’s what you need to know:

• What’s an Agent?

  Agents are systems designed to accomplish objectives. They become more powerful when equipped with reasoning, decision-making, tool usage, and context awareness.

• What’s a Workflow?

  Workflows define the steps needed to achieve a goal—like gathering requirements, designing, implementing, testing, and deploying.

• Agents + Workflows = Smarter Systems

  Workflows don’t require agents—but agents can supercharge them. With reasoning, tools, and context, agents can optimize workflows. This is the foundation of multi-agent systems, where multiple agents collaborate to tackle complex objectives.

• Meet Microsoft Agent Framework

  Microsoft Agent Framework is a powerful set of .NET libraries that makes building and orchestrating AI agents easier than ever. From simple chatbots to complex multi-agent workflows, it streamlines development, deployment, and monitoring with minimal boilerplate code.

• Built on Proven Foundations

  The framework builds on trusted technologies like Semantic Kernel, AutoGen, and Microsoft.Extensions.AI—bringing reliable, production-ready tools to developers. Learn how to build your own agent here.

Want to bring AI-powered Agent actions into your Power Apps and Dynamics 365 such as analyzing data, generating insights, updating records, and surfacing recommendations? Explore Infinity AI Agents and contact us today.

Visual Studio September 2025 Update Brings Powerful AI-Driven Enhancements

Microsoft has rolled out the September 2025 Visual Studio update, introducing a fully integrated agentic AI experience. From code generation and modernization to profiling and code reviews, these features are designed to transform the way developers build and optimize applications.

Here’s what’s new:

• Profiler Agent

  No need to be a profiling expert—Profiler Agent uses AI to analyze CPU usage, memory, and runtime behavior. It highlights performance bottlenecks, creates or optimizes BenchmarkDotNet benchmarks, and suggests fixes in a guided loop—all seamlessly integrated with GitHub Copilot.

• .NET Application Modernization

  Upgrade your .NET apps to the latest versions and migrate to Azure faster with the Copilot-powered modernization agent, designed to streamline upgrades and migrations.

• MCP Prompts, Resources & Sampling

  Model Context Protocol (MCP) introduces a standardized way for servers to expose structured prompts, contextual resources, and agentic behaviors—making LLMs more powerful in development environments.

• Mermaid Diagram Generation

  Visual Studio now supports rendering Mermaid diagrams directly in the Markdown editor—whether you write the syntax yourself or let Copilot generate it for you.

• Smart Code Reviews

  GitHub Copilot takes local code reviews to the next level with AI-driven insights that make reviews faster, smarter, and more effective.

Transform the way your organization works. Learn how a custom application can be tailored to your needs—contact us today.

Visual Studio’s New MCP Features Aim to Deepen Copilot Context Awareness

Microsoft announced that Visual Studio will now support MCP prompts, resources, and sampling, bringing richer context from engineering stacks directly into the IDE. These enhancements are designed to upgrade the Copilot experience by enabling more intelligent, context-aware developer workflows.

Here’s what you need to know:

• Why Use MCP?

  Seamlessly integrate external tools and services to make your development workflow more intelligent and context-aware.

• MCP Resources & Templates

  Model Context Protocol (MCP) lets servers share files, database schemas, or app-specific info—each identified by a URI—so you can pull relevant context directly in the IDE.

  How to use: In chat, click +Reference → MCP Resource, choose a template, fill arguments, and add the resource.

• MCP Prompts & Prompt Templates

  Provide structured instructions tailored to your server’s content for sharper, task-specific responses.

  How to use: Click +Reference → Prompts → MCP Prompts to select and apply.

• MCP Sampling

  Enable advanced agentic behaviors where MCP servers can make nested LLM calls, unlocking multi-step, automated tasks—Visual Studio handles this automatically if your MCP supports it.

Curious how chatbots and automated workflows come to life? Explore our NZPA case study and get in touch with us today.

Microsoft Launches All-in-One Marketplace for Cloud, AI, and Business Apps

Microsoft has officially launched Microsoft Marketplace, a unified platform that brings together Azure Marketplace and AppSource into a single destination to discover, purchase, and deploy cloud and AI solutions. The new Marketplace features tens of thousands of solutions, integrates AI apps and agents, and aims to simplify how customers and partners engage with Microsoft’s cloud ecosystem.

• What Are Frontier Firms?

  They’re leaders accelerating AI transformation to enrich employee experiences, reinvent customer engagement, reshape processes, and unlock creativity and innovation.

• How to Get Started

  The Microsoft Marketplace is your trusted hub for cloud solutions, AI apps, and intelligent agents. It extends the Microsoft Cloud, bringing partner innovations to customers worldwide.

• Why Marketplace Matters

  • AI Apps & Agents for Any Need

    Access thousands of AI-driven solutions to automate tasks, speed decision-making, and boost business value—all designed to work seamlessly with your existing Microsoft stack.

  • Comprehensive Catalog

    Explore solutions spanning data, analytics, productivity, collaboration, and industry-specific scenarios—integrated with Azure, Microsoft 365, Dynamics 365, Power Platform, Microsoft Security, and more.

  • Integrated Experience

    Discover, deploy, and use partner-built solutions directly inside Microsoft products for a smooth, secure workflow.

Want to harness the power of AI to uncover insights, update records, and more in your Dynamics 365 and Power Platform? Explore Infinity AI Agents and contact us today.

Power BI Brings Modeling Tools Directly to the Browser

Microsoft has rolled out general availability of semantic model editing directly in the Power BI service, enabling users to modify models through their web browsers for the first time. This update broadens the platform’s capabilities by bringing core modeling workflows—once limited to desktop tools—into the online experience.

Here’s what you need to know:

• Create new semantic models in the web

  Connect to 100+ data sources directly in your browser. Go to the Create page, select Get data, choose your source, configure settings, and select the tables to include—then pick your next step.

• Option 1: Create a report

  Jump right into the familiar report editor and start building visuals, adding filters, and customizing layouts—no desktop app required.

• Option 2: Only create a semantic model

  Focus on modeling first. Add relationships, write DAX measures, and configure row-level security before creating reports.

• Option 3: Transform data

  Use the modern Power Query editor to shape and transform your data in the browser, then decide whether to create a report or just a semantic model.

• Edit existing semantic models in the web

  Manage relationships, author DAX, edit properties, configure row-level security, and refresh schema and data—all online.

Want to see Power BI in action? Read our IPCA case study to discover how they transformed complaints handling, created dashboards for trend analysis, and gained full transparency with Power BI integration.

Looking to explore Power BI? Contact us today for expert guidance.

Microsoft Announces General Availability of TMDL View in Power BI

Microsoft has officially released TMDL view as generally available in Power BI, moving it out of preview as of September 2025. The new release gives users full control and visibility over semantic models through a modern, code-first editing experience.

Here’s what’s new:

• Enhanced Code Editor

  Enjoy syntax highlighting for DAX and Power Query, helpful tooltips, auto-formatting, and quick code actions—similar to the TMDL Visual Studio Code extension.

• TMDL Diff View

  Review and compare changes before applying them for safer, more transparent updates.

• Smarter Usability

  Drag-and-drop sections and intuitive UI upgrades make editing faster and easier.

• Stronger Robustness

  Power BI Desktop is now hardened to open and edit any semantic model, enabling seamless collaboration with external tools, code files, or AI-powered edits.

Ready to turn data into powerful insights? Explore Business Intelligence and connect with our team today.

Power BI Reports in Teams Now Open in Standalone Windows

Microsoft is rolling out a change in Teams so that Power BI content shared in chat no longer takes over the entire interface. Instead, opening a Power BI item launches it in its own pane beside the chat — letting users keep conversations visible while exploring reports.

• How it works

  When someone shares a report (or any Power BI item) in a Teams chat or channel, it now opens in a separate window. Your original chat stays visible in a collapsible side panel.

• Why it matters

  View the chat alongside the report to keep the conversation going while exploring your data. Continue using Teams—switching between chats, channels, or apps—without losing your place. If multiple preview cards are shared, the same window updates with the latest item.

• What it means for you

  Seamless collaboration and multitasking. Stay in flow, stay connected—all inside Teams

Microsoft Power BI Introduces Flexible Calendar-Based Time Intelligence in Preview

Microsoft Power BI has rolled out a preview of calendar-based time intelligence, enabling organizations to use custom calendar schemas — whether retail, fiscal, lunar, or otherwise — for time-based analytics. Previously, Power BI’s time intelligence functions were constrained by assumptions like continuous date tables and support mainly for Gregorian calendar structures.

Here’s what you need to know:

• What are Calendars?

  Calendars define which columns represent specific time attributes, enabling fully customized time-based analysis. You can even define multiple calendars on any table for complete control over how time is segmented and analyzed.

• Key Benefits

  Works with any calendar type, makes no structural assumptions, supports sparse dates, introduces new week-based calculations, and delivers noticeable performance gains.

• Getting Started

  Enable the preview by navigating to Options & Settings > Options > Preview Features, then turn on Enhanced DAX Time Intelligence. Next, define calendars using either the Calendar Options UI or the TMDL view.

• Using the Calendar Options UI

  Open the Table Tools tab or right-click a table in the Data pane to view existing calendars, create new ones, or access the Mark as Date Table options.

• Using TMDL View

  Enjoy more flexibility, including tagging columns as time-related for advanced scenarios.

Copilot Gets Smarter: Microsoft Lays Out How to Use Prompt & Instruction Files

Microsoft has published guidelines for two new file-types — Instruction Files and Prompt Files — designed to help developers give Copilot clear rules and context when generating code. Learn what are they, when to use them, and how to use them.

Instruction Files

• What are they?

  Instruction files often named copilot-instructions.md defines the rules, coding standards, or guidelines Copilot should follow when generating code in your repo or workspace.

• When to use?

  Perfect for team projects or open-source repos to keep Copilot’s suggestions aligned with org standards and naming conventions.

• How to use?

  Create your file from a sample or scratch. Save it in .github/ or add file-type-specific instructions in .github/instructions/. Copilot automatically applies them to every request.

Prompt Files

• What are they?

  Prompt files provide context for a specific coding session or file, steering Copilot’s output for a targeted task.

• When to use?

  Great for prototyping, custom algorithms, integrations, compliance requirements, or onboarding new team members needing explicit guidance.

• How to use?

  Draft your prompt and place it in .github/prompts/ as [promptname].prompt.md. In VS Code type /[promptName] in chat, or in Visual Studio type #[promptName] to activate.

Microsoft Unveils Private Preview: Azure Boards Now Integrated with GitHub Copilot

Microsoft has opened a private preview that connects Azure Boards with GitHub Copilot, allowing developers to move work items seamlessly into AI-assisted coding. This integration aims to accelerate planning and development while keeping project tracking in one streamlined workflow.

Here’s what you need to know:

• Tasks you can automate:

  You can do bug fixes, incremental features, test coverage improvements, documentation updates, and technical-debt cleanup.

• Step 1 – Prepare the work item:

  Write clear instructions in the work item description, then click “Create a pull request with GitHub Copilot.” Select your repository and add any extra details.

• Step 2 – Copilot gets to work:

  The Copilot coding agent generates a branch and draft pull request, automatically linking it back to the Azure Boards item.

• Step 3 – Review the results:

  Depending on complexity, Copilot completes the plan and updates the work item status once the pull request is ready for review.

• Requirements to enable:

  To use this feature, you must have GitHub-hosted repositories, enable the Azure Boards–GitHub integration, activate the Copilot agent on those repositories, and have your organization approved for the private preview. Make sure to update the Azure Boards GitHub App permissions if needed.