How risk scores are calculated in Asset Risk Predictor

Every day, our AI algorithm users your aggregated sensor data to assign risk scores (and corresponding risk levels) to the sensors and assets you've included in Asset Risk Predictor.

Our AI algorithm uses 4 main steps to calculate risks:

• Step 1: Training to establish a baseline
• Step 2: Generating risk predictions
• Step 3: Assigning risk scores
• Step 4: Assigning risk levels

These steps are outlined in the sections below.

Step 1: Establishing a baseline

When you first purchase Asset Risk Predictor, our AI algorithms are automatically trained for 7 days to establish a baseline for your assets. After 7 days, once Asset Risk Predictor starts generating predictions, it compares the current sensor values to this baseline to help identify risks.

Step 2: Generating risk predictions

Our algorithm uses aggregated sensor data to generate risk predictions per sensor, per minute. These predictions are generated using 3 different calculation methods, which inform the risk score assigned into the next step.

Step 3: Assigning risk scores

Next, the algorithm averages the results from the 3 calculation methods in the step above. Using this average allows for more reliable, accurate predictions than just using one calculation method.

It then uses this value to assign hourly and daily risk scores between 0 - 1000.

Step 4: Assigning risk levels

Lastly, assets and sensors are then assigned a risk level based on the risk score they were assigned in the previous step.

By default, risk levels are assigned as follows:

However, your organization can change how risk levels are defined. To learn more, see Change risk level definitions for Asset Risk Predictor.