GitHub Copilot Testing for .NET Now Generally Available in Visual Studio 2026

Microsoft has officially rolled out GitHub Copilot testing for .NET as part of Visual Studio 2026 v18.3, bringing AI-driven unit test generation directly into the IDE. The new capability helps developers automatically generate, build, and run high-quality tests for C# code, shortening the feedback loop between writing and validating code. Early users can now leverage richer IDE integration and more natural prompting for widespread test automation.

Here’s what you need to know:

• Built for real testing workflows

  The testing agent understands your solution structure, frameworks, and build system — acting as an end-to-end workflow, not just a one-off prompt.

• Natural, free-form prompting

  Simply start with @Test and describe what you want in plain language — no rigid commands required.

• More control over test generation

  Generate or fix tests, target coverage goals, and apply your team’s testing conventions using simple prompts. You can reference specific code areas or even include current Git changes for more precise results.

• Getting started

  Open your C# project or solution, start a new Copilot Chat thread, and begin your prompt with @Test, then select Send. Test Explorer shows the results as the tests are generated, and once complete, GitHub Copilot testing provides a summary in Copilot Chat.

Complement your Dynamics 365 CRM with custom business applications for maximum productivity. Contact us today.

Power Apps Introduces MCP and Enhanced Agent Feed in Public Preview

Microsoft has introduced a new public preview for Power Apps, highlighting its growing focus on AI-assisted business applications. The update brings MCP and an enhanced agent feed, pointing toward more interactive and intelligent user experiences. It represents another step in turning everyday apps into collaborative workspaces between people and AI.

Here’s what this means for organisations:

• Automate repetitive app tasks

  AI agents can extract data from emails, files, or shared folders and create records automatically — with human approval before completion. Less manual entry, fewer errors, faster processing.

• Human-in-the-loop supervision

  Users can review, guide, or take control of agent actions directly inside the app through built-in supervision tools. Keep governance and trust while benefiting from automation.

• A redesigned collaborative agent feed

  A shared workspace where humans and AI work together — with side-by-side comparisons, task tracking, and performance insights. Clear visibility into what AI is doing and why.

• Turn existing apps into agentic apps

  Embed agents into the apps employees already use daily — no rebuild required. Faster adoption and real productivity gains.

• Scalable AI for real business processes

  From data entry to broader CRUD operations, the MCP framework enables enterprise-scale automation across workflows — assistance → autonomy → collaboration.

Want AI to help with your work in Dynamics 365 and Power Apps? Explore Infinity AI Agents and contact us today!

Dynamics 365 Contact Center Sets New Standards for AI Evaluation

Microsoft is highlighting a shift in how organisations approach AI in customer service, emphasizing that evaluation is no longer just a technical step but a core business capability. In its latest guidance for Microsoft Dynamics 365 Contact Center users, the company outlines key dimensions organisations must consider to ensure AI systems remain reliable and customer-focused. The message is clear: trustworthy AI in contact centers depends on continuous, multi-stage evaluation — not a one-time test.

Here are the key facets every organisation should consider:

• Development-stage vs Production-stage evaluation

  Test early in development to catch issues before deployment, and evaluate in production to understand how AI performs under real-world traffic and customer behavior.

• Manual vs Automated evaluation

  Use automated tests for scalable, continuous checks (e.g., after deployments or config changes), and pair them with manual evaluations for deep insights, context, and judgment that machines can miss.

• Platform vs Customer-run evaluations

  Internal evaluations from the AI provider establish a baseline of quality and safety, while customer-run tests reveal real usage patterns and constraints specific to your business.

• Synthetic data vs Production traffic

  Start with synthetic or simulated data for safe, repeatable tests (especially edge cases), then validate against real production traffic to gauge how AI behaves with actual user interactions.

• Evaluation after vs during execution

  Post-interaction analysis unlocks deep learning about correctness and resolution quality, while real-time evaluation lets you intervene instantly (e.g., escalate to a human when sentiment drops).

Ready to create AI agents in Dynamics 365 and Power Apps that work the way you do? Explore Infinity AI Agents and get in touch with us today.

Power BI Phases Out Old Excel & CSV Import Experience

Microsoft has announced upcoming changes to the Power BI service that will affect how some users work with Excel and CSV data. A legacy data import experience is being retired, with key deadlines arriving in 2026. Users relying on older reports are encouraged to transition to the newer data connection methods to avoid outdated data.

Here’s what you need to know:

• What’s affected?

  Semantic models created with the old import experience can’t be edited in the browser, downloaded, or scheduled for refresh. If your model supports those capabilities, you’re already using the newer connectors and won’t be impacted.

• What’s changing?

  Starting July 31, 2026, reports built using the old import experience will stop refreshing from their source files. They will still open and be editable, but the data will no longer stay synced unless recreated using the new connectors.

• Key dates to remember

  By May 31, 2026, creating new semantic models using the old experience will be disabled. By July 31, 2026, existing models built with it will stop refreshing.

Unlock smarter insights with Power BI — explore Business Intelligence and get in touch today.

Microsoft Improves Large-Codebase Navigation in Visual Studio

Microsoft is enhancing AI-assisted development in Visual Studio with the introduction of the find_symbol tool. The update gives AI agents richer, language-specific context so they can better understand relationships across a codebase instead of relying on plain text search. It’s part of Microsoft’s continued push to make Copilot more reliable inside real-world projects.

Here’s why it matters:

• What is find_symbol?

  find_symbol gives Copilot deep, language-aware insight into your code. It understands functions, classes, interfaces, and variables, including their declarations, implementations, scope, and every reference across the codebase—so the agent can reason, not just search.

• Example #1: Enhancing existing functionality

  Need to add logging, telemetry, or performance metrics without breaking anything? Copilot can quickly identify all related symbols and references, helping ensure updates are complete and accurate.

• Example #2: API refactoring

  Hardening or refactoring an API requires knowing exactly how it’s used. With symbol-level insight, Copilot can trace call paths, understand dependencies, and suggest safer refactors with minimal risk.

Go beyond CRM — streamline your operations with custom business applications. Talk to us today.

Dynamics 365 Contact Center Introduces Proactive Voice Engagements

Microsoft has announced new proactive voice engagement capabilities for Dynamics 365 Contact Center, continuing its push to modernize customer service operations. The update focuses on helping organizations reach customers more efficiently and consistently through outbound communication scenarios. It marks another step in strengthening enterprise-grade engagement tools within the Dynamics 365 service ecosystem.

Here’s what you need to know:

• Why this matters

  Proactive voice scenarios become far more consistent and reliable. Your team spends less time guessing and more time actually helping customers.

• Answering Machine Detection (AMD)

  If the system detects voicemail, it automatically follows the flow you set — play a message, end the call, or continue the journey. No more wasted agent time on non-productive connections.

• SIP-based call outcomes

  Every outbound call is automatically classified (LiveAnswer, AnsweringMachine, Busy, NoAnswer, Invalid Address, etc.). You instantly know what happened — and what to do next — without manual tagging or extra configuration.

• Predictive dial mode

  Calls are placed ahead of agent availability based on real-time metrics like wait time and abandonment rate. The result: customers connect right when an agent becomes free, improving both efficiency and experience.

• What’s next?

  Microsoft is expanding channel coverage and dialing performance — including conversational SMS, preview dialing improvements, and native campaign management.

Strengthen customer relationships from first touch to long-term loyalty with a smarter CRM. Discover how it works and talk to us today!

Microsoft Brings Code Apps to Power Apps for Broader Development Scenarios

Microsoft has announced the general availability of code apps in Power Apps, marking a major step in expanding the platform’s development model. The update reflects Microsoft’s push to bring professional development and low-code experiences closer together within the Power Platform. It also aims to help organizations manage custom applications under a more unified environment.

Here’s what you need to know:

• What are Code Apps?

  Code Apps bring the full power of the Power Platform to web developers. Build in your preferred code-first IDE, iterate locally, and run the same app seamlessly inside Power Apps using modern frameworks while keeping full control over your UI and business logic.

• For developers: stay in your flow

  Develop using familiar frameworks and local workflows, then deploy directly to Power Platform with access to 1,400+ connectors and enterprise-ready hosting.

• For IT: control without friction

  Every Code App is automatically governed—enforcing authentication, DLP policies, conditional access, monitoring, and lifecycle management without slowing developers down.

Turn your processes into powerful custom apps. Let’s build it together.

.NET Framework 3.5 Moves to Standalone Deployment in Future Windows Versions

Microsoft is changing how legacy components are delivered in upcoming Windows releases, signaling a gradual shift in platform maintenance strategy. The update affects a long-standing dependency still used by many older enterprise applications and reflects broader lifecycle planning across the Windows and .NET ecosystem. Developers and IT teams are now being encouraged to prepare for modernization as the platform evolves.

Here’s what you need to know:

• What’s changing?

  Starting with Windows 11 Insider Preview Build 27965, .NET Framework 3.5 will no longer be included as an optional Windows feature. Instead, it must be installed separately using a standalone installer.

• Who is affected?

  The change applies to new Windows platform releases moving forward. It does not impact Windows 10 or existing Windows 11 releases up to 25H2.

• Why the shift?

  With .NET Framework 3.5 reaching end of support on January 9, 2029, Microsoft is encouraging organizations to start planning migrations to supported .NET versions.

• What should you do?

  If your application depends on .NET Framework 3.5, Microsoft has published detailed guidance — including installers, compatibility notes, and recommended migration paths.

Enhance your Dynamics 365 and Power Platform with custom apps built around your workflow. Let’s talk today!

Microsoft Reveals February AI Roadmap for Visual Studio

Microsoft has shared its latest look ahead at how artificial intelligence will evolve inside Visual Studio. The February roadmap signals a shift toward stabilizing AI workflows while continuing to expand the role of agents in everyday development. The update hints at deeper integration and smarter interactions across the IDE experience.

Here’s what developers like you can expect:

• Agent Mode & Coding Agents

  Improved progress indicators for long-running tasks, smarter failure handling and retries, plus enhanced diagnostics and logging — helping agents recover gracefully and making issues easier to troubleshoot.

• Planning Agent

  A new dedicated experience designed for scoping, refining, and iterating tasks before execution — moving beyond the lightweight planning mode developers use today.

• Copilot SDK & Platform Integration

  Early work begins to integrate the Copilot CLI directly into Visual Studio Copilot for a more connected development workflow.

• Model Context Protocol (MCP)

  Continued investment in secure and scalable tool connectivity, including Client ID metadata documents and governance support for MCP servers inside Visual Studio.

• Models & Context Management

  Under-the-hood improvements to keep Copilot fast even with large contexts — simplifying context reduction and preventing unexpected behavior.

• Copilot Experience in the Editor

  Smoother day-to-day coding with improved acceptance shortcuts, reduced IntelliSense conflicts, and better quick-info support.

Boost productivity across Dynamics 365 and the Power Platform with AI. Discover Infinity AI Agents — talk to us today.

Microsoft Adds Custom GitHub Copilot Agents to Azure Boards Workflow

Microsoft has announced an update to the Azure Boards integration with GitHub Copilot that introduces support for custom AI agents. The enhancement allows development teams to leverage tailored Copilot agents directly within Azure Boards, streamlining the transition from work items to automated coding tasks.

Here’s what you need to know:

• What are custom agents?

  Custom agents are tailored versions of the GitHub Copilot coding agent, designed to follow your team’s workflows, coding standards, and tool preferences. Think of them as specialized teammates that consistently apply your rules—no need to repeat instructions every time.

• Creating custom agents

  You can define a custom Copilot coding agent by creating a .agent.md profile file in your GitHub repository. This file contains the agent’s instructions, tools, and behavior, making it easy to align Copilot with specific tasks or workflows.

• Using custom agents in Azure DevOps

  Once created at the repository or organization level, your custom agent automatically becomes available in Azure DevOps. When creating a pull request from a work item, you’ll see a new agent selection option alongside the repository list—making it seamless to choose the right agent for the job.

Boost productivity in Dynamics 365 and Power Apps with custom AI agents — explore Infinity AI Agents and get in touch with us today!