Act on the event
while it is still happening.

Real-time intelligence on Microsoft Fabric turns operational events into live KPIs, alerts and actions while the outcome can still change. Eventstream ingests the events, Eventhouse stores them and serves them through KQL, a live KPI board shows the condition, and Activator fires the scoped alerts and actions. The useful measure is how quickly the team recognizes the condition and responds, not query speed alone.

A live pulse is our two-week, fixed-scope first slice: one stream family on your own events, taken to a live KPI board. It runs on Pulse, our real-time accelerator, which covers ingestion, live KPIs, anomaly alerts and automated action. You keep the KPI board, and the working slice becomes the basis for choosing the next event family and the next action.

A business needs real-time data when the time gained changes an operating decision. When an order falls behind, a machine changes behavior or a vehicle sends an exception, a refreshed report can arrive after the useful window to intervene has narrowed. Where no decision changes with faster information, the reporting cycle is enough. We start from that window and work back to the stream.

Query latency is how fast the data answers; decision latency is how fast the team can act on what it shows. A fast query is one part of the path from event to action. We time the whole path, from the event landing to the alert reaching the team or system that acts on it, against the team’s decision window.

You need one event stream and the decision it should change. The first working session names the event family, the KPI and the intervention it should support, and that decision bounds the two-week live pulse. Further implementation, with more event families and more actions, is then scoped from the working first slice rather than from a plan on paper.