Skip to main content

Sentiment Feedback

Sentiment Feedback (Genesys Cloud Quality & Performance Management)

Section Description
Module Context Part of Quality Management → Speech & Text Analytics within Workforce Engagement Management (WEM)
Admin Location Admin → Quality → Sentiment Feedback
Purpose Allows administrators to correct sentiment classification errors in Speech & Text Analytics by labeling phrases as Positive, Neutral, or Negative, improving automated sentiment detection accuracy

Sentiment Feedback is used to improve the machine learning sentiment model that evaluates transcripts produced by Speech & Text Analytics. When the system misinterprets emotional tone in a conversation, administrators can add a phrase and assign the correct sentiment classification.

These corrections influence future sentiment analysis results and help improve topic detection, interaction insights, and analytics reporting.

Summary Table

Attribute Details
Feature Area Speech & Text Analytics
Primary Function Improve sentiment classification accuracy
Data Source Interaction transcripts
Supported Channels Voice (transcribed), Chat, Messaging, Email
Sentiment Labels Positive / Neutral / Negative
Dialect Context Sentiment phrases are tied to a language dialect used by speech analytics models
Typical Users Quality administrators, analytics administrators
Dependency Speech & Text Analytics must be enabled

Study Notes

Topic Explanation
Sentiment Analysis Automatic classification of emotional tone within conversations
Sentiment Feedback Manual correction of sentiment detection errors
Phrase-Based Overrides Administrators add phrases that override default sentiment classification
Transcript Dependency Sentiment feedback only applies to interactions with transcripts
Dialect Awareness Phrases must match the speech analytics dialect model used during transcription
Analytics Impact Changes influence future sentiment scoring and analytics dashboards

Key concepts:

  • Sentiment feedback improves the accuracy of automated sentiment scoring.
  • Phrases are applied per dialect, ensuring correct language interpretation.
  • Corrections apply to future transcript processing.
  • Sentiment feedback helps refine analytics used for CX insights and coaching.

Navigation

Task Navigation
Open Sentiment Feedback Admin → Quality → Sentiment Feedback
Add new sentiment phrase Admin → Quality → Sentiment Feedback → Add Phrase
Edit phrase Admin → Quality → Sentiment Feedback → Edit
Delete phrase Admin → Quality → Sentiment Feedback → Delete
Review transcript sentiment Performance → Workspace → Interactions

Configuration Fields (UI Form Fields)

Main Page

UI Field Description Options / Behavior
Phrase List Displays phrases configured for sentiment feedback Read-only
Sentiment Label Shows sentiment classification assigned to phrase Positive / Neutral / Negative
Dialect Dialect the phrase applies to Example: English – United States
Created By User who created phrase entry Read-only
Created Date Timestamp of phrase creation Read-only
Search Search for phrases in list Text field
Add Phrase Opens phrase creation form Button
Edit Edit phrase sentiment configuration Button
Delete Remove phrase from sentiment list Button
Refresh Reload phrase list Button

Create/Edit Form

UI Field Description Options
Phrase Phrase appearing in transcripts Text input
Sentiment Sentiment assigned to phrase Positive / Neutral / Negative
Dialect Speech analytics dialect model used for phrase interpretation Example options: English – United States / English – United Kingdom / Spanish – Spain / Spanish – Mexico / Portuguese – Brazil
Description Optional note describing reason for classification Text input
Save Save configuration Button
Cancel Discard changes Button

Tabs, Toggles, Dropdowns, Action Buttons

UI Element Type Item
Tabs Sentiment Feedback page
Text Fields Phrase, Description
Dropdowns Sentiment classification, Dialect
Buttons Add Phrase, Save, Cancel, Edit, Delete, Refresh

Note: No toggles are documented for Sentiment Feedback in the admin UI.

Dependencies

Component Purpose
Speech & Text Analytics Generates transcripts used for sentiment analysis
Interaction Recording Required for voice transcript generation
Topics Sentiment can be correlated with topics
Programs Topics grouped for analytics programs
Topic Miner Detects candidate phrases for sentiment feedback
Language Models Dialect-specific transcription models

Platform Integration / Related Components

Component Relationship
Speech & Text Analytics Provides transcript and sentiment analysis engine
Topics Sentiment often analyzed alongside topics
Programs Topic groups used in analytics dashboards
Topic Miner Identifies phrases administrators may classify
Interaction Analytics Displays sentiment metrics

Integration Examples

Integration Description
Analytics API Retrieve sentiment metrics for dashboards
Notifications API Subscribe to interaction lifecycle events to trigger external analytics processing
Conversations API Retrieve transcript data programmatically

Example integration workflow:


Customer Interaction
↓
Speech Analytics Transcript
↓
Sentiment Analysis
↓
Sentiment Feedback Overrides
↓
Analytics API retrieves results
↓
External dashboard updated

Related Topics / Further Reading

Topic Description
Speech & Text Analytics Core analytics engine
Topic Miner Discover phrases in transcripts
Topics Phrase detection logic
Programs Topic grouping for analytics
Evaluation Forms Agent performance scoring

Implementation Checklist

Task Status
Enable Speech & Text Analytics ☐
Confirm transcript generation ☐
Verify correct dialect configuration ☐
Review sentiment accuracy ☐
Identify misclassified phrases ☐
Add sentiment feedback entries ☐
Validate improved sentiment detection ☐

Implementation Guide

Step Action
Step 1 Enable Speech & Text Analytics
Step 2 Confirm transcripts are generated
Step 3 Review interactions for sentiment misclassification
Step 4 Navigate to Sentiment Feedback
Step 5 Click Add Phrase
Step 6 Enter phrase and select correct dialect
Step 7 Assign sentiment classification
Step 8 Save configuration

How to Implement

Phase Description
Analytics Monitoring Monitor sentiment results in transcripts
Error Identification Detect incorrect sentiment classification
Phrase Feedback Add phrase classification using correct dialect
Validation Monitor analytics improvements

Workflow


Customer Interaction
↓
Interaction Recording
↓
Speech-to-Text Transcription
↓
Speech Analytics Sentiment Engine
↓
Admin Detects Incorrect Sentiment
↓
Sentiment Feedback Phrase Added (Dialect-specific)
↓
Future transcripts classified correctly

Architecture Diagram


Customer Interaction
↓
Recording Engine
↓
Speech-to-Text Processing
↓
Transcript Storage
↓
Speech & Text Analytics
├ Sentiment Analysis
└ Topic Detection
↓
Sentiment Feedback Phrase Overrides
↓
Analytics Dashboard

Real Flow Scenarios

Scenario 1 – Correcting Positive Sentiment


Customer says: "Thank you for resolving my issue"
↓
System labels phrase as Neutral
↓
Admin adds phrase with Positive sentiment
↓
Future interactions labeled Positive

Scenario 2 – Detecting Customer Frustration


Customer says: "This service is terrible"
↓
Transcript generated
↓
Sentiment engine detects negative tone
↓
Analytics dashboard flags interaction

Usage Scenarios

Scenario Description
Customer satisfaction monitoring Improve sentiment accuracy
Product issue detection Identify negative feedback
Agent coaching Understand customer reactions
CX analytics Track sentiment trends

Implementation Examples

Example Configuration
Support center Add satisfaction phrases
Billing queue Label complaint phrases
Sales queue Label positive purchase indicators

Design Example


Inbound Customer Call
↓
Speech Analytics Transcript
↓
Sentiment Classification
↓
Admin Reviews Transcript
↓
Sentiment Feedback Added
↓
Improved Analytics Accuracy

Best Practices

Practice Reason
Add phrases tied to correct dialect Prevent incorrect language interpretation
Focus on high-frequency phrases Improve model quickly
Review sentiment dashboards regularly Detect classification issues
Combine sentiment with topics Improve context understanding

Naming Convention

Resource Example
Phrase Classification Positive_Service_Resolution
Phrase Classification Negative_Billing_Issue

Naming pattern:


<Sentiment>_<Context>

Security Considerations

Control Description
Role-based access Restrict configuration to administrators
Transcript privacy Protect sensitive conversation data
Data retention policies Control transcript storage
Encryption Protect stored transcripts

Limitations / Constraints

Constraint Description
Sentiment categories Positive / Neutral / Negative
Requires transcripts Speech analytics must be enabled
Context sensitivity Phrase sentiment may vary by context
Dialect dependence Phrase must match configured dialect

Troubleshooting

Issue Cause Resolution
Sentiment incorrect Phrase not defined Add sentiment feedback
Phrase not matching Dialect mismatch Select correct dialect
No transcripts available Speech analytics disabled Enable transcription
Analytics unchanged Phrase not used in transcripts Verify phrase frequency

Interview Cheat Sheet

Question Answer
What is Sentiment Feedback? Manual correction of sentiment classification
Where is it configured? Admin → Quality → Sentiment Feedback
What labels exist? Positive, Neutral, Negative
What additional parameter is required? Dialect
What dependency is required? Speech & Text Analytics

Key Takeaways

Topic Summary
Sentiment Feedback Improves automated sentiment detection
Phrase Overrides Admins classify phrases manually
Dialect Support Phrases must match speech analytics dialect
Speech Analytics Provides transcripts used for analysis
Continuous Improvement Feedback improves analytics accuracy

Screenshots

No Comments
Back to top