Intelligent Alerts overview

Last update: 2024-01-19
  • Topics:
  • Alerts
    View more on this topic
  • Created for:
  • User

Intelligent Alerts allows for more granular control over alerts and integrates anomaly detection with the alert system.

Here is a video tutorial on Intelligent Alerts (5:34)


Intelligent Alerts let you:

  • Build alerts based on anomalies (90%, 95%, 99%, 99.75%, and 99.9% thresholds; % change; above/below)
  • Preview how often an alert will trigger
  • Send alerts by e-mail or SMS with links to auto-generated Analysis Workspace projects
  • Create “stacked” alerts that capture multiple metrics in a single alert

There are three ways get to the Alert Builder:

Method Details
Go directly to the Alert Builder Components > Alerts
Use the keyboard shortcut in Workspace Ctrl + Shift + A (Windows) or Cmd + Shift + A (Mac)
Select one or more freeform table line item/s Right-click and select Create Alert from Selection. This opens the Alert Builder and pre-populates the appropriate metrics and filters applied from the table. You can edit the alert if needed. Create alert from selection

The percent thresholds are standard deviations. For example, 95% = 2 standard deviations and 99% = 3 standard deviations. Depending on the time granularity you choose, different models are used to calculate how far away (how many standard deviations) each data point is from the norm. If you set a lower threshold (such as 90%), you get more anomalies than if you set a higher threshold (99.75%).


Using timestamped data to create alerts can cause alerts to fire incorrectly. Adobe recommends using non-timestamped data for Intelligent Alerts.

Anomaly lookback for alerts

If an alert uses anomaly detection, the training period varies based on the granularity selected for the alert.

  • Monthly granularity: 15 months + same range last year
  • Weekly granularity: 15 weeks + same range last year
  • Daily granularity: 35 days + same range last year
  • Hourly granularity: 336 hours

See Statistical techniques used in Anomaly Detection for more information.

On this page