Success stories

How mobilezone gave a four-person IT team a proactive morning shift

Three narrowly scoped AI agents help a Swiss retailer catch IT issues early and prepare orders, while employees keep control of final decisions.

Small IT team reviewing store alerts before opening

What happened

Swiss telecom retailer mobilezone deployed three Microsoft Copilot Studio agents across IT monitoring, B2B order entry and product creation, according to a Microsoft customer story. The company reported CHF 274 million in 2025 net sales from continuing operations in its annual report.

The clearest use case is the morning IT check. A four-person support team uses a monitoring agent to identify and resolve an average of one to two issues before 125 stores open. Microsoft reports that initial support response times fell to within 30 minutes, compared with backlogs that could previously last three days to a week.

A separate sales-order agent reads customer PDFs and creates draft orders, saving five to ten minutes per order. Employees still review and confirm every transaction. When the system finds an ambiguous field, it sends the question back to a person rather than guessing.

The reported results come from the vendor’s customer case, not an independent audit. mobilezone’s investor reports provide additional company context, but the supplied material does not establish a broader independently verified return on investment.

Why it matters

The useful lesson is not that a retailer replaced its support team. It gave a small team bounded detection and preparation work at operational choke points.

That distinction matters for owners. A four-person IT function cannot be everywhere at once, but it may be able to start each day with a prioritized list of issues instead of waiting for stores to report them. Likewise, preparing an order can be worthwhile even when approval must remain with an employee.

A practical pattern emerges:

  • Start with a narrow, frequent workflow where delays are visible.
  • Let the system flag uncertainty instead of filling gaps with guesses.
  • Keep people responsible for consequential approvals.
  • Measure response time, minutes saved and prevented disruption—not agent count.
  • Require each proposed use case to show its time, cost and expected value before development begins.

For a small business, the unit economics are straightforward: reclaimed minutes and avoided disruption must outweigh setup, integration, monitoring and review costs. The strongest first project may be a proactive morning shift, not a general-purpose chatbot.

Sources

Let's talk

Wondering where AI fits in your business?

Tell me what you're working on and where things feel stuck.