Innovate Your Business
with Our Insights!

The first version focuses on requests. Supervisors can list pending time-off, shift swap, and shift bid requests across their planning groups, filter them by type, status, and date range, and approve or reject time-off requests inside the same conversation, with scheduling changes applied to WEM records in real time. Representatives see only their own submissions, so a question like “What’s the status of my shift swap request?” returns their data and nothing else. 

Security follows the model teams already run. The tools register in the Service Agent tool catalog, run through the Microsoft Agent 365 Tooling Gateway, and enforce Dataverse role-based access control at the MCP layer, so no extra permission setup is needed. Administrators switch tools on by security role, team, or pilot group from the Dynamics 365 Contact Center admin center, which lets a rollout start with one planning group before it widens. 

Microsoft has two more versions planned. Version 2 adds a retrieve-schedule tool that shows representatives their upcoming shifts, queue assignments, and break windows, and gives supervisors coverage views by queue and time slot. Version 3 is planned to add live leave balances from connected HR systems such as Workday, plus clock in and clock out from any Microsoft 365 surface. 

For service leaders, the bigger signal is that MCP is becoming the standard way to put business actions inside the tools people already use, a shift we break down in our explainer on the Model Context Protocol for Dynamics 365. The tools only work as well as the queue and planning group structure behind them, so it helps to review how your omnichannel customer service setup is organized first.

Microsoft positions Activate for three scenarios: 

For now, Salesforce is the only supported source system. 

Microsoft is also making the preview available to implementation partners. Partners can continue using their existing migration methodologies while using Activate to support parts of the discovery and assessment process. 

Early users have focused mainly on that discovery capability. Comcast used Activate to gain greater visibility into its Salesforce environment and identify important migration considerations. Ziwi Pets has also highlighted the additional visibility and validation the tool provided during delivery. 

Microsoft says future additions will include agentic mapping and test sampling, natural-language migration guidance, and executive assessment summaries. 

The significance of Activate is not that it removes the need for migration planning. 

It gives teams more information earlier. 

Instead of beginning with a largely manual discovery exercise, migration teams can use the tool to surface data structures, dependencies, and customizations before making major design decisions. 

That can help answer a more important question than simply, “What can we migrate?” 

It can help teams decide what should migrate, what should be simplified, and what should be redesigned for Dynamics 365. 

Activate is still a public-preview tool, though, and its current scope is limited to Salesforce migrations. Broader decisions around architecture, integrations, data quality, security, testing, change management, and user adoption still sit outside the tool itself. 

The change supports organizations that want to keep Dynamics 365 as the commerce engine without being tied to a standard storefront. Microsoft’s composable architecture separates the experience layer from the underlying commerce logic, allowing businesses to build their own customer-facing channels while continuing to use Dynamics 365 for capabilities such as pricing, inventory, loyalty, checkout, orders, returns, and fulfillment. 

At the center of the model is Commerce Scale Unit, which exposes these commerce capabilities through RESTful APIs. 

Microsoft describes two headless models. 

Headless orchestration allows a custom or third-party application to call Commerce Scale Unit APIs and use Dynamics 365 Commerce during the transaction itself. This can support custom storefronts, mobile applications, marketplaces, and other purpose-built experiences. 

Headless import is designed for scenarios where the transaction happens in an external system first and is then brought into Dynamics 365 Commerce for downstream processing. 

Microsoft says its new licensing options support both approaches. Pricing was not published in the announcement, so organizations will still need to confirm licensing terms before choosing an architecture. 

The company highlighted fashion retailer PETER HAHN as one example of the composable approach. The retailer combines a custom web shop with Dynamics 365 Commerce modules and specialized third-party solutions while connecting commerce activity with its finance and supply chain processes. 

Microsoft is also positioning the architecture for a future in which commerce interactions extend beyond traditional websites and apps. AI-powered agents and conversational experiences could introduce new customer, employee, and B2B channels without requiring businesses to rebuild the commerce logic underneath each one. 

For retailers considering a custom or headless experience, the update makes licensing part of the architecture decision itself. The key question is no longer only how the storefront should look, but also where the transaction should happen and what role Dynamics 365 Commerce should play behind it.

The update shows Microsoft moving ERP further toward AI-assisted and agent-driven processes, but organizations will need to separate what is usable today from what is still on the roadmap before building project timelines around it. 

In Finance, Microsoft is expanding automation across journals, reconciliation, accounts payable, collections, variance analysis, and business intelligence. The Account Reconciliation Agent is being developed to resolve exceptions in bulk, while AI invoice recognition and finance-entry suggestions are aimed at increasing touchless processing. 

Supply Chain Management receives the broadest set of additions. Procurement Agent impact analysis is in public preview and can trace supplier-proposed purchase order changes to affected sales, production, transfer, and inventory plans. Demand planning generative insights are already generally available, while Supplier Engagement, dynamic warehouse item placement, and rental management are at different stages of preview or roadmap development. 

Commerce updates focus on B2B buying, Trade POS, Tap to Pay, faster checkout, and the Commerce MCP server, which is currently in public preview. Project Operations is also adding change order management, subscription billing, revenue recognition enhancements, and a time-phased staffing view. 

The broader direction is clear: Microsoft wants agents to work across ERP processes, business data, and collaboration tools. For customers, the immediate task is to map each capability by release status before it influences scope, budgets, or go-live dates.

Work IQ entered public preview on September 30, 2026, with rollout continuing through October and Dynamics 365 ERP integration expected later in the month. 

Work IQ is built around three components. Semantic models let organizations connect and model business data once so multiple agents can work from the same business context. Business skills package organizational instructions and supporting resources into reusable assets, reducing the need to maintain separate prompts for each agent. Governed actions allow agents to recommend or make changes to source records while following user permissions and approval requirements. 

The service runs on Microsoft Dataverse and also introduces developer tools including the Dataverse Ask API, a Work IQ MCP server, command-line tools, and support for coding agents. Governance is handled through Microsoft Agent 365, with administrators controlling which environments, data, and AI policies can participate. 

Work IQ covers Dynamics 365 sales and service applications, Finance and Operations apps, Power Apps, and third-party business applications, with usage billed through Copilot Credits. 

The broader significance is that Microsoft is trying to solve one of the harder problems in enterprise AI: giving multiple agents consistent access to business context, permissions, and rules. That also raises the importance of data quality, because shared semantic models will carry source-data errors into every agent that relies on them.