A machine learning system may monitor usage of network or hosted resources by users or employees, and may dynamically generate short or “micro” surveys for immediate presentation. These surveys may be aggregated and analyzed by the machine learning system, reducing delays of responses. As a result, engagement may be measured in real-time or near real-time, actionable insights generated, and responsive actions taken. The machine learning system may monitor various interactions of users or employees with a virtual or hosted environment or workspace, including connections to virtual machines, remote desktop applications, SaaS applications, web applications, or other such entities, as well as environmental characteristics such as network location and/or quality.