Overview

When exporting the Sentiment Signal Report or in the Console case view, users may observe that multiple signals (for example, multiple follow-up requests) are detected within a single inbound comment. This is an expected behaviour of the SupportLogic's Machine Learning (ML) model.

Explanation

SupportLogic’s ML model analyzes customer comments at the sentence level.
Each sentence within a comment is independently processed to identify potential signals such as Follow-Up Request, Escalation Risk, Frustration, and others.

Because of this design:

• A single comment containing multiple sentences can generate more than one detected signal.

• These signals may be identical (e.g., two Follow-Up Requests) or different (e.g., a Follow-Up Request and a Frustration signal).

This approach enables more granular detection and helps capture nuanced customer intent within longer messages.

Example

Inbound Comment Example:

“Can you please check on the previous issue again? Also, I haven’t received any update from the team.”

Detected Signals:

• Follow-Up Request (Sentence 1)

• Follow-Up Request (Sentence 2)

Both signals appear within a single comment since each sentence is analyzed individually.

Summary

Multiple signals within a single comment are a designed behaviour of the SupportLogic's ML model, ensuring fine-grained sentiment and intent detection at the sentence level rather than at the comment level.