Microsoft Tames Virtual Machine Support Tickets with AI-Powered Troubleshooting Agent

Microsoft's IT organization, Microsoft Digital, has developed an AI-powered solution to automate the process of resolving repetitive support tickets for virtual machines (VMs) in its Managed Cloud Labs platform. The platform hosts over 150,000 VMs and allows employees to create and manage lab environments as needed.

Before implementing the AI solution, engineers would spend significant time manually troubleshooting issues with VMs, which could take hours to resolve. However, with the new Troubleshooting Agent, a human support engineer can select a VM and report an issue, and the agent will investigate, diagnose the cause, and attempt a repair in minutes.

The Troubleshooting Agent is integrated with Azure Serial Console and uses AI to streamline the troubleshooting process. It provides a running summary of what it's detecting and the actions it's taking, making the automated work visible to the user but also 'hands-free'. If the agent can't safely resolve an issue, it provides the full diagnostic context to the user for human support engineer intervention.

The solution has been a great success in reducing wait times for common issues, which once took over 90 minutes to resolve. With the Troubleshooting Agent, these issues are now identified and resolved in less than 10 minutes, providing immediate self-remediation and reducing downtime.