AI–Powered Real-Time Intelligence with Anomaly Detection (Preview)
A new AI capability in Microsoft Fabric’s Real-Time Intelligence (RTI) is available with a preview of anomaly detection. This marks the start of a journey to empower users with AI-driven, scalable, proactive real-time data experiences.
Learn more about the other Real-Time Intelligence announcements in our documentation.
Unlocking Value with AI in Real-Time Intelligence
We are witnessing the rise of AI-powered systems that perceive, reason, act, and learn in continuous feedback loops. RTI is the infrastructure layer that powers them. It connects and makes sense of signals across all different sources (time and space, physical and digital) and can scale to the size and speed of it.
Real-Time Intelligence is designed to help users make sense of streaming data and act on it instantly. Anomaly detection helps users uncover ‘unknown unknowns’ in their data. Whether it’s spotting irregularities in sensor data or identifying operational bottlenecks, anomaly detection uses industry-standard algorithms to recommend the best models for your data. It democratizes complex data science tasks, making them accessible to less technical personas and enabling organizations to act on insights faster.
Together, this offers proactive intelligence that drives operational efficiency, reduces risk, and enhances decision-making.
How to Use Anomaly Detection
Real-time data often follows predictable patterns and trends, and spotting when it deviates from this is a way to derive high value from your data. Previously, anomaly detection has been the realm of data scientists and needed code to be written. Our new anomaly detection feature brings this to all users of Real-Time Hub.
To use Anomaly Detection, start by selecting an Eventhouse in Real-Time Hub. You select the ID fields and values that you want to look for anomalies in and let the system analyze it. The detector has a library of industry-standard models which it tests to find which is best at predicting the patterns in your data.
You’ll find a preview of any anomalies that were discovered, and you can try the different models to confirm which captures the right anomalies in your data. You can publish those detected anomalies as Anomaly events to the Real-time Hub and set alerts on them to automatically get Teams messages and emails when an anomaly is detected.
Enable these features in your Fabric tenant and give us your feedback!
Anomaly Detection is available to try out now! You’ll need to enable it in your Fabric admin portal. Then, you can create detector configurations from the Real-Time hub.
Please let us know your feedback!
You can contact us in comments for this blog post, or through feedback form.
