Microsoft has introduced a managed MCP server in Dynamics 365 Commerce to ensure AI agents interact securely with retail systems.
Without a secure layer between AI agents and retail systems, AI can become a significant operational and security risk and organizations lose more control over its actions.
CX leaders must therefore ensure that their AI agents are able to complex customer-facing tasks whilst maintaining the basic governance and security required for reliable experiences.
Gaining an early preview of the MCP server, Shaun Morley, Group Retail and Omnichannel Systems Manager at Michael Hill, highlighted that retailers can accelerate AI adoption by building on their existing commerce infrastructure.
“Our early experience with testing Dynamics 365 Commerce MCP server has been promising,” he noted.
“Having access to this type of tooling creates a practical way to introduce AI-driven innovation at the pace retail demands, simply by layering intelligence over existing commerce platforms.”
Retail’s Growing Security Blind Spot
Today’s AI assistants can already recommend products or summarize policies, however rising expectations means that organizations are instilling more involvement in the customer journey, meaning more direct access to core commerce systems.
Without secure access, retailers are required to build custom integrations for every AI application, resulting in complexity and increased operational risk.
When an agent is not operating within strict authorization in retail, it could perform unauthorized transactions or be manipulated through malicious prompts that can quickly become customer trust issues when payment information and sensitive data is involved.
Furthermore, retailers may adopt multiple AI assistants all without a common framework, creating fragmented governance and making it difficult to monitor AI activity.
Without standardized, secure interfaces, retailers risk accumulating dozens of custom connections and isolated governance practices, increasing enterprise maintenance costs and expanding the attack surface.
Closing the Gap Between AI and Retail Systems
Announcing the public preview at NRF 2026, Microsoft revealed Dynamics 365 Commerce MCP Server, a new capability that enables AI agents to securely interact with Dynamics 365 Commerce through a standardized interface.
The next generation of agentic commerce now means that AI agents can execute more complex tasks with live commerce data, with MCPs now emerging as an industry standard that defines how AI agents communicate with enterprise applications.
The server exposes approved commerce capabilities as standardized tools that any MCP-compatible AI agent can use, meaning organizations can connect compatible platforms to Dynamics 365 Commerce without building custom integrations for each one.
The MCP Server aims to solve the challenges within retail by acting as a secure, managed gateway between AI agents and retail systems, where authorization and governance remain under enterprise control and AI agents only receive permitted capabilities.
By positioning Dynamics 365 Commerce as an AI-ready commerce platform, Microsoft is preparing retailers for a multiple AI agent operation by connecting each one through the same governed interface.
AI agents can therefore securely perform transaction and support employees using natural language, creating opportunities for AI to become an active participant in customer journeys.
Gudni Vilmundarson, Director at Amicis Solutions, explained how the MCP Server can reduce complexity for frontline employees while improving operational efficiency.
“Getting early access to the Dynamics Commerce MCP server has been a fantastic experience for us as a Microsoft partner,” he said.
“We used it to build a voice-driven commerce agent that puts natural-language intelligence right on top of Dynamics 365 Store Commerce, so store associates can search the catalog, manage carts, look up customers, and handle everyday POS tasks just by talking.”
Securing the AI-Retail Divide
For CX, AI that can complete customer journeys is becoming a valuable capability for customers that need to switch channels to complete transactions, with retailers able to complete a customer’s list of queries and needs through a single conversation.
However, as AI becomes more capable, governance is becoming just as important as intelligence, as Heath Ramsey, GVP of Product Management, AI Platform at ServiceNow, told CX Today:
“The promise of AI is the value. Promise is the transformation, but the gap is the governance and the visibility, especially within the CX space. Because it’s such an important part to make sure that you get the experience right the first time.”
This reduces friction and improves the quality of AI-powered experiences by building trust through consistent responses grounded in real-time business data over disconnected databases, a common critical driver of customer trust.
For CX leaders, the announcement signals that AI governance is becoming just as important as the capability, where success will increasingly depend on enterprises creating a secure architecture for AI to perform transactions safely.
As a result, leaders will need to consider how AI fits into the end-to-end customer journey, how permissions and governance are managed, and how customer data is protected while still enabling personalized, efficient experiences.
As conversational interfaces become a new engagement channel for commerce, designing experiences will require AI effectively accessing enterprise system and executing transactions to deliver accurate outcomes.
Those that treat AI as a standalone chat interface may face obstacles when meeting growing customer expectations for AI that can reliably get things done.
