mobilezone redefines productivity with Microsoft Copilot Studio agents

Building an in-house agent capability

mobilezone is a Swiss telecommunications retailer and service provider with 125 stores across Switzerland. Its stores depend on 183 point-of-sale (POS) devices and support high volumes of daily customer interactions while back-office operations support hundreds of B2B customers.

By combining Microsoft Copilot Studio, with Dynamics 365 Commerce and Microsoft Dynamics 365 Finance mobilezone established a foundation for building agents across its retail and business operations.

mobilezone created three agents to address friction in two distinct business areas. On the retail side, the Store Monitoring Agent helps a four-person IT service management team oversee store devices. On the B2B side, the Sales Order Creation Agent and Product Creation Agent address repetitive work in back-office operations and product catalog management.

Executive sponsorship has helped mobilezone move beyond experimentation and establish a repeatable internal capability. The company gives engineers room to test ideas, learn from early deployments, and apply lessons to new business processes. Mark Schwarz, Chief Digital and Information Officer, mobilezone explains, “We see agentic AI as a strategic capability that reshapes how work gets done. By building expertise in-house with Microsoft Copilot Studio, we can scale innovation faster, reduce complexity, and create sustainable business value across the organization.”

Reactive support and repetitive work

Before implementing the Store Monitoring Agent, mobilezone’s support team managed issues across each of its 183 POS devices. Store employees typically responded to problems by creating support tickets and waiting for Field Support Engineer Marino Kneringer and his team to identify the root cause.

Application errors and hardware-station faults were difficult to diagnose and often required escalation across multiple teams. As a result, resolving issues took longer, and some problems were not identified until they had already disrupted transactions and affected customers. "Problems surfaced only after they had already impacted transactions and customers, which in retail and sales is a business risk," said Kneringer.

The scale of operations made proactive oversight difficult. Kneringer and three first-level supporters cover 125 stores and 183 POS devices, while also handling broader hardware and office support responsibilities. Monitoring fleet health and identifying isolated device errors created noise and consumed valuable time for the small support team.

AI and Automation Engineer, mobilezone, Fabian Kistler faced a different challenge on the B2B side. Employees processed high volumes of sales orders based on standard customer contracts. Previously, back-office workers had to transfer information manually from order documents into the ERP system, a repetitive process that took several minutes per order.

Although the retail and ERP challenges were different, mobilezone identified a common pattern. The retail scenario needed better visibility into physical devices and store operations, while the ERP scenarios required controlled automation that could read business documents, create records, and preserve human review.

Three agents for distinct business workflows

Despite limited AI experience, Kneringer adapted a prebuilt Copilot Studio Store Monitoring Agent and deployed it in only three weeks. Microsoft Azure Monitor agents collect selected logs from every POS device, all of which run on Dynamics 365 Commerce, and Azure resources. These signals are processed by the Copilot Studio agent.

Previously, when an issue occurred, such as a cash register failing to open, the support team reacted by remotely accessing the affected device and analyzing raw system outputs to identify the cause. But now, each morning at 7:00 AM, the agent posts a Microsoft Teams update, highlighting devices with the highest error volumes, error categories, and the most recent occurrence of each issue. Support specialists can then ask natural-language questions about specific devices and receive AI-generated summaries of complex technical information.

The result is a more proactive support process. Kneringer can narrow a 24-hour report to the most relevant three- to six-hour window, investigate unusual error patterns, and address issues before stores open at 9 a.m. As he explains, “I start at 7:00 in the morning, while the stores open at 9:00. I query all the issues that happened overnight and resolve them before the stores open.”

For the B2B business, Kistler rebuilt a sales order agent in Copilot Studio and connected it to the Dynamics 365 ERP environment through the Microsoft ERP connector and a Model Context Protocol (MCP) connection. An employee drops a customer PDF into the agent through Microsoft Teams. The agent extracts the order information, maps it to the correct fields, creates a draft sales order, and posts status or warning messages for the back-office group. A person reviews the draft and confirms it in Dynamics 365, preserving human control over the final transaction.

Earlier fixes and time returned to employees

Store support teams now begin each day with actionable information instead of waiting for complaints. Kneringer reports resolving an average of one to two issues each morning before they have a negative impact on business. As part of broader support-process improvements, support tickets now receive an initial response within 30 minutes, compared with previous backlogs that sometimes delayed response by three days to a week.

The Sales Order Creation Agent is returning measurable time savings for employees. According to Kistler, "For each contract or order that we enter, we are saving anywhere from 5 to 10 minutes." For its first automated customer, mobilezone processes approximately 30 to 40 orders each month, saving five to ten minutes per order, saving hours of manual data entry each month. The agent works for several minutes while the employee handles email or another task, then the employee checks the draft and confirms it.

The human-in-the-loop design also supports trust during the early production period, with particular attention to quantity and price. The agent flags missing or unclear information and can leave fields empty for review rather than presenting uncertain data as complete.

Expanding proven patterns responsibly

mobilezone plans to extend the Sales Order Creation Agent to additional B2B customers after validating customer-specific document formats. The company is also exploring a process that allows business groups to submit agent ideas and estimate time savings, cost savings, and business value so leaders can prioritize the next opportunities.

The Product Creation Agent is expected to move into production once the remaining requirements are completed.

For store operations, Kneringer is testing the next version of the Store Monitoring Agent. Potential enhancements include alert-based monitoring, connections to the company's knowledge base, automatic ticket creation, and guided remediation. He also wants store managers to receive clear, natural-language summaries. mobilezone is taking a measured approach because more frequent Azure queries and autonomous actions may change the cost profile, and higher-impact fixes will continue to require appropriate human oversight.

Learn more about mobilezone by following them on LinkedIn, Facebook, Instagram, and Tik Tok.

Looking for more inspiring agentic AI transformation stories? Explore additional customer successes on the Microsoft Agent Transformation Stories site.