Innovate Your Business
with Our Insights!

The server ships with more than 90 service-oriented tools covering the full breadth of the service workflow. AI clients can list, open, summarize, update, reassign, and resolve cases, manage queues, and track SLAs. They can pull customer context in the form of account and contact details, activity timelines, and case summaries, and they can work the knowledge base end to end, from searching articles and answering questions in natural language to drafting new content, identifying knowledge gaps, and publishing approved materials. Email handling, next-best-action recommendations, and direct Dataverse queries and updates round out the set. 

Because the server is built on the Model Context Protocol, organizations expose these capabilities once and make them available across every supported AI experience, rather than building and maintaining a custom integration for each client. Supported clients at launch include Microsoft 365 Copilot, Microsoft Copilot Studio agents, Visual Studio Code, GitHub Copilot CLI, ChatGPT, Claude Code, and other HTTP-based MCP-compatible clients. 

Governance is handled through the Microsoft-hosted Agent 365 Tooling Gateway, with access aligned to existing Dataverse roles and permissions, meaning an AI client can only see and change what the signed-in user is already entitled to. Administrators can also register additional external MCP servers and connect custom Copilot Studio agents, leaving room to extend the model well beyond Microsoft’s out-of-the-box tools. 

The release builds on a public preview launched last winter and lands a month after Service Agent in Microsoft 365 Copilot reached general availability, cementing MCP as the connective standard for Microsoft’s agentic customer experience stack. 

The deadlines give the guidance its weight. Extended support for Dynamics NAV ends in January 2028, and Service Plan coverage and subscription licenses end on April 30, 2031, after which no additional perpetual licenses can be purchased. For Dynamics GP and Dynamics SL, Microsoft reiterated the timelines announced in June 2025, with product availability ending April 30, 2031 for GP and January 15, 2030 for SL. 

The recommended destination is Dynamics 365 Business Central, which Microsoft positions as a modern cloud ERP spanning finance, operations, inventory, sales, and projects. The pitch to legacy customers centers on built-in AI capabilities, continuous updates that eliminate disruptive upgrade cycles, deep integration with Microsoft 365, Teams, Power Platform, and Azure, and reduced IT complexity once infrastructure moves to the cloud. 

Microsoft outlines a five-stage path to get there: assess the current environment and customizations, prepare the upgrade path, migrate and validate data, configure roles and train users, and execute the cloud transition. Comparison guides, video tutorials, and documentation mapping legacy capabilities to Business Central accompany the announcement. 

The framing is deliberate — a natural evolution rather than an emergency measure — but the clock is running, and customizations, data volume, and integration complexity all take time to work through. Organizations weighing their options can find practical, experience-based guidance in our Dynamics NAV to Business Central upgrade guide and our breakdown of what Dynamics GP end of life means for your timeline.

The most consequential of the three is the Journey Creation Agent, which builds launch-ready campaigns from natural language prompts and eliminates manual journey setup. It is joined by the Outreach Optimization Agent, which automatically tailors timing and follow-up sequences to each customer’s behavior while respecting consent preferences, and the Customer Insights MCP Server, which lets AI agents across applications act on customer data while honoring privacy requirements in real time. 

The reasoning behind the repositioning is Microsoft’s argument that AI is only as effective as the data behind it. Customer Insights supplies that foundation through two core elements: Customer Insights Data, a unified customer profile that consolidates fragmented signals into a continuously updated single source of truth, and Customer Insights Journeys, AI-powered orchestration that turns those insights into personalized customer engagement across channels. 

Microsoft describes the change as helping organizations move “from a legacy system of record to become a system of action.” For marketing teams, the direction of travel is clear: campaign orchestration is becoming agent-driven, a shift we have been tracking in marketing automation for Dynamics 365, where journey design, segmentation, and outreach timing increasingly run on AI rather than manual configuration.

The feature answers the question that stalls most AI rollouts in service organizations: how do we know the AI is making the right decisions? Until now, teams had to rely on configuration reviews and theoretical validation before switching automation on. Shadow Mode replaces that with evidence. The agent runs in parallel with human workflows, observing genuine incoming cases and simulating the actions it would take, while the system displays each recommendation and the reasoning behind it in real time. No data is modified, and complete operational isolation is maintained throughout. 

Validation spans the agent’s three main jobs. For case enrichment, Shadow Mode evaluates how accurately the agent predicts routing and prioritization fields. For case resolution, it examines how the agent identifies intent and recommends solutions. And for follow-up and closure, it tests the agent’s lifecycle decisions around timing and eligibility. 

Setup is minimal: administrators open the Copilot Service admin center, navigate to Case settings and then Case Management Agent, and activate Shadow Mode for the capability they want to observe. 

The release extends the validation story Microsoft began with case enrichment simulation earlier this year, and it reflects a broader pattern across AI-powered case management in Dynamics 365: trust is built by letting teams verify AI behavior against their own real-world data before granting it autonomy.

The two biggest time-savers deal with bulk changes. A new Move To capability lets schedulers select multiple bookings at once — Ctrl+click on PC, Cmd+click on Mac — and shift them together by a consistent timeframe, instead of updating each booking individually when plans change. Its companion, Reassign To, lets schedulers select multiple bookings and hand them to a different resource through a single streamlined interface, covering the moments when a technician becomes unavailable or work needs consolidating. 

The board’s views are also expanding. A dedicated Map View now sits alongside the existing List and Gantt views, giving geography-dependent operations a cleaner, more focused way to work with location-based scheduling. The board can additionally display week numbers alongside dates, so teams that plan against week-based calendars no longer convert dates manually. 

The fifth enhancement, partial cancellation, addresses long-running engagements: multi-day bookings can now be modified by cancelling only specific segments while preserving the remainder, removing a workaround that complex projects previously required. 

None of these are headline AI features, and that is the point. Microsoft framed the release as listening to scheduler feedback and strengthening core functionality — the unglamorous work that determines whether dispatch teams actually hit their SLAs. For a full picture of what the platform can do beyond the board, see our A–Z guide to Dynamics 365 Field Service capabilities.