How does Elsa’s Sentiment analysis work? (ELSA) (Practera 1)

Learn about how Elsa detects negative sentiment in any submissions by students and mentors and which action she suggests.

Detecting Participant and Mentor Sentiment

Whenever you login to your Practera dashboard, Elsa goes through each moderated, reflection, survey and team360 assessment type submission and review/feedback and scans the content for negative sentiment. Only submission in the past 7 days are analysed.

To be considered in the analysis, text fields need to have more than 20 characters.

All text answer fields in a single submission are analysed together. Should negative sentiment get detected, Elsa creates a todo list entry.

The todo list entry will be visible for 7 days. If you haven’t actioned the suggestion, Elsa will assume that it was not relevant and remove it from your interface.

Suggested action

Elsa will prompt you to send a message to the respective people. When clicking on the todo list item, the message pop up window will appear.

Before sending a message, you can examine the sections that Elsa has determined as negative by clicking on the “eye” icon next to a person’s name.

If you think Elsa has wrongly detected negative sentiment, you can simply ‘uncheck’ the box next to the respective name.

Note: Elsa is constantly learning how to become better at detecting and predicting issues. She is far from perfect, so sometimes she will recommend an action that is irrelevant. 

Available message templates

The are currently no specific templates available.

Tracking

Elsa tracks your sentiment interventions automatically and informs you whether or not a participant has actioned your suggestion. Learn how