Negative Sentiments Workflow

Modified on Fri, 19 Jul at 9:56 AM

Purpose

Reduce escalations: identify cases where there is potential negative feedback from a customer, and to determine if action is required to improve the customer experience.

Summary

Open SupportLogic and go to: OPERATIONS > Console > Negative Sentiments and follow the instructions below.


Negative Sentiments Workflow

Frequency

Once per day. Some customers with heavy case volume may need to complete twice per day.
Time Commitment

15  - 20 minutes
Audience

Support Managers and Support Team Leads
Goal

Daily negative sentiment acknowledgment at 90%
Product Area

Console




The Console is a command and control center that surfaces cases to you in real-time based on Natural Language Processing and Machine Learning. 


  1. Navigate to the Console by selecting it from the navigation bar on the left.
  2. Select your relevant Global Filter so that you see only the cases that are important to you and/or fall under your purview.
  3. Set your date range appropriately based on your needs:
  • Last 7 days - do this at the start of your week so you see any cases that surfaced during the weekend or that were not addressed the prior week.
  • Since Yesterday - do this at the start of your day to see any relevant cases that came in since the end of your prior workday.
  • Today - do this throughout your day to see anything new that has surfaced during your shift.

Negative Sentiments tab

  1. Select the Negative Sentiment Tile on the top navigation bar
  2. The cases shown will be those with negative sentiment detected during the selected time frame
  3. Set the signals filter to show signals detected on Inbound comments only
  4. Group the sentiments by any number of options, though we recommend using
  • Signal Type
  • Elapsed Time
  • Sentiment Score

Review Sentiments Detected

As seen in the animation above, you have several options to review and either acknowledge or correct the sentiments detected in the case text:


  • You could click the check box on the tile for a particular case to acknowledge the sentiment detected
  • Open the case, review the sentiments detected at the top of the view
  • Navigate to the specific location within the text to review the sentiment in context


Continue to review sentiments detected and move onto the next case.



Correcting SupportLogic When Wrong Sentiments Are Detected

Depending on your permissions, you can correct the ML engine of SupportLogic by either adding a sentiment that you detect within case text or by correcting a sentiment that was incorrectly identified by SupportLogic.


To add a sentiment that was missed by SupportLogic:

  1. Open the Support Hub view of the case
  2. Review the case text and highlight the sentiment that was missed
  3. Mouse over the highlighted text and select Label
  4. Navigate through the hierarchy of sentiments and select all that apply


To correct a sentiment mis-labeled by SupportLogic:

  1. Open the Support Hub view of the case
  2. Review the detected sentiments to identify the one that is incorrect
  3. Click the ellipsis next to the sentiment label and then Change label
  4. Locate the incorrectly identified sentiments and uncheck it
  5. Select Apply


The animation below shows you how to both Add sentiments and how to remove sentiments to a case in SupportLogic:


Sentiment labeling and correcting labels


Sentiment Feedback

Feedback provided by you is collected and reviewed to update the ML engine for SupportLogic and updates the sentiment(s) identified within that case text for your organization.




Additional Resources

Below are two videos that step through how to complete the Negative Sentiments workflow and review your organization's performance is trending regarding detected sentiments.


Review the Decrease Negative Sentiments Workflow Video here


Learn how to find Negative Sentiments in your UI here.



Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article