Create Smart Alerts for a Listening use case

Updated 

Let us dive into a sample use case corresponding to the Listening module and see how a Smart Alert is configured for the same.

As a Brand Manager at a consumer goods company, you want to get alerted on any trend shifts on –

  • Mentions fetched from your Brand Topic.

  • Only in the English Language.

  • Interested in only the mentions categorized as negative sentiment.

  • Get alerted on Volume and Engagement anomalies detected by the algorithm.

  • Send out an email only on anomalies where the mentions are more than 100.

  • You want to further narrow down your alert emails only to highly severe anomalies. (A proxy for the same could be more than 7 on a rating of 0-10).

  • Send mail to a list of recipients including those who have access to Sprinklr and other external folks

  • You want to add a custom line to the mail according to the configuration details

  • You want to include timestamps in every alert mail

  • To avoid spam in your inbox, you want to limit the number of alerts to a maximum of 25 in a single day.

Steps to create a Smart Alert for this use case

  1. In the alert configuration window, enter the name of the alert and description/tag (optional).

  2. Select Listening as the alert module. Smart Alerts is pre-selected as the alert category in the form.

  3. Under the Alert dimensions section, add the topic corresponding to your brand.

    Topic containing BackfillCheck. You have now defined the dimension you want to get alerted for.

  4. Now, add a few more filters as per your requirement –

    • Click Add New Filter and select Sentiment as the dimension. Then select Positive and Negative as values.

    • Click Add New Filter and select Language as the dimension. Then select English as the value.

  5. By default, the Alert on all entities combined option will be selected as the entity combination type. For more details on the logic for this section, refer to the Choose entity combinations section in the Listening Smart Alerts article.

  6. Choose Engagement Alert and Volume Alert from the dropdown in the Smart Alert Type field. You have now requested Sprinklr to store two separate time series (Engagement and Volume) for the set of Alert dimension filters​.

  7. In the Choose time range aggregation for Smart alerts anomaly detection, both the checkboxes are selected by default. This will ensure that alerts are triggered on both kinds of aggregations.

  8. Avoid spam by adding maximum notification frequency, add Maximum alerts as 25, and set the Duration as Daily.

  9. Click Next to go to the Alert distribution details tab. Refer to this section – Alert distribution details in the main article to understand what details you need to fill in here.

  10. In the Refine Alerts e-mail based on other parameters section, add the following filters –

    • Choose Mentions from the dropdown and set it as greater than 100

    • Choose Alert Severity from the dropdown and set it as greater than 7

  11. Click Save. Wait for a few minutes, and your alert will be up and running to monitor any anomalies

You have successfully created an alert to monitor trends in Volume and Engagement around your brand’s mentions.