Today, we’re excited to publicly announce the DataVisor Automated Rules Engine, a rules engine that maintains itself. By eliminating the need for human effort to add, tune, and delete rules, it’s the most sophisticated rules engine on the planet.

At a glance, the Automated Rules Engine:

  • adds, deletes, and tunes rules
  • backtests all rules against historical data
  • provides visualizations and metrics to monitor rule health

…all without any manual effort.


Infusing rules with the latest machine learning technology

When we talked to people that work with rules engines, we discovered a love-hate relationship. On the one hand, rules engines are great because they’re easy to understand, modify, and deploy. On the other hand, rules quickly decay, and they’re straightforward for fraudsters and criminals to get around. Maintaining the hundreds or thousands of rules within the rules engine also takes a lot of time.

Many of the problems with a rules-based system are solved by using unsupervised machine learning. Unsupervised machine learning catches evolving attacks automatically before any damage is conducted, without labels. The Automated Rules Engine uses the results of our Unsupervised Machine Learning Engine and creates human-readable rules around them. As a result, the Automated Rules Engine automatically adapts to evolving attacks and deprecates rules that are no longer effective.

Asset 5@4x

A full stack analytics platform

The Automated Rules Engine is our latest addition to our full stack analytics platform. Each detection method–rules, supervised machine learning, and unsupervised machine learning–has its own benefits and drawbacks. By using them all together, you can enjoy the benefits and mitigate the weaknesses of each approach.


DataVisor’s flagship technology is the Unsupervised Machine Learning Engine, a detection method that automatically adapts with evolving attacks, without labels. This works hand-in-hand with our Supervised Machine Learning Engine, which takes in labels from the Unsupervised Engine and can also ingest outside labels. The Supervised Engine is good at finding individual attackers that are similar to historical known bad actors.

Now, we’ve added the Automated Rules Engine, which automatically learns from the Unsupervised Engine as well. The Automated Rules Engine infuses the power of the latest machine learning and big data technologies into the familiar and extensible rules engine.

The new Automated Rules Engine is part of Version 3.0 of the DataVisor User Analytics Service,  available now for all DataVisor customers. Learn more about how our Automated Rules Engine can dramatically reduce manual rule maintenance time and decrease the window of threat exposure, or contact our team to get personalized information about how we can help.