Most dashboards tell you what happened yesterday. A few tell you what happened this morning. Almost none tell you what is happening right now, while there’s still time to do something about it.
That gap between data arriving and someone actually acting on it is exactly what Real-Time Intelligence in Microsoft Fabric was built to close.
For years, most Fabric workloads have assumed you can wait. Data lands in a lakehouse or warehouse overnight, a pipeline runs, a semantic model refreshes, and someone checks a report the next morning. That rhythm works fine for monthly revenue trends. It falls apart when you need to catch a fraud pattern forming right now, a machine drifting out of spec on a production line, or a delivery fleet slipping behind schedule.
Real-Time Intelligence is Fabric’s answer to that second kind of problem. It isn’t one tool. It’s a connected set of them, purpose built for data that keeps moving instead of sitting still in a table.
What Real-Time Intelligence Actually Covers
Real-Time Intelligence is the part of Fabric dedicated to data in motion: streaming events, logs, IoT signals, and anything else that shows up continuously instead of in a nightly batch. Everything that flows through it lands in the Real-Time hub, a catalog view of every stream in your tenant, so you can see what’s available without hunting through individual workspaces. The same governance and permission model that covers your lakehouses and warehouses applies to your streams too.
The Building Blocks
A handful of items do most of the work, and each one has a distinct job.
Eventstreams
- Bring streaming data in without writing ingestion code, using connectors for sources like Apache Kafka clusters, database change data capture feeds, Amazon Kinesis, Google Cloud Pub/Sub, MQTT, and Microsoft’s own real-time weather connector
- Filter, clean, transform, and route events before they land anywhere, using windowed aggregations and content based routing
- Set up alerts directly on the stream, without leaving the Eventstream canvas
Eventhouse
- A database built specifically for streaming, semi-structured, and free text data, tuned to query billions of events in seconds
- Acts as a container that can hold multiple KQL databases, which share capacity so you aren’t standing up separate infrastructure for every project
- Comes with unified monitoring across every database it holds
KQL Querysets and Real-Time Dashboards
This is where you actually explore what landed in the Eventhouse and turn it into something visual. Dashboards built on top of an Eventhouse can refresh continuously instead of waiting on a scheduled job, which matters when the whole point is seeing changes as they happen.
Activator
Activator is the no-code rules engine that sits at the end of the chain. It watches a stream or a query result, and the moment a condition is met, it fires an action: an email, a Teams message, a Power Automate flow, or a call out to another system. For straightforward stateless rules, that reaction happens in subsecond time.
A Useful Way to Picture It
Think about a city traffic control room. Sensors along every road feed in constant readings. That’s the Eventstream layer, quietly collecting and cleaning data before it goes anywhere. All of that history lands in a central data room built to hold and query enormous volumes of it fast: the Eventhouse. Controllers watch live screens built from that data, which is the Real-Time Dashboard layer. And when a sensor shows gridlock forming at a specific intersection, an alarm goes off automatically instead of waiting for someone to notice. That’s Activator doing its job. None of those pieces is impressive on its own. Wired together, they’re the difference between reacting to a problem and preventing one.
Where It Meets the Rest of Fabric
Real-Time Intelligence doesn’t replace your lakehouse or warehouse. It complements them. You can query an Eventhouse the same way you’d explore other data in Fabric, and Power BI reports can connect to that same real-time data instead of waiting on a scheduled refresh. If you’ve read our post on Power BI storage modes, think of Real-Time Intelligence as covering the case those modes weren’t really built for: data that’s still arriving while someone is looking at the report.
Why It Matters
Plenty of organizations collect real-time data today and still treat it like batch data by the time anyone looks at it. The value of Real-Time Intelligence isn’t speed for its own sake. It’s closing the distance between an event happening and a person or a system being able to respond to it. As more of Fabric’s AI and automation features start reacting to live conditions instead of static reports, that gap is exactly where the next round of useful features will get built.