Back to Generative AI Features Quality AI uses LLM-powered features to analyze agent–customer conversations. It evaluates agent performance, detects sentiment, identifies compliance gaps, and extracts actionable insights in real time and after the interaction.
Model Feature Matrix This table lists Quality AI features and their supported model providers and models.
(✅ Supported | ❌ Not supported)
Feature Azure OpenAI OpenAI Gemini Custom LLM Advanced Topic Discovery GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Agent Empathy Identification GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ By Value Adherence validation for Quality AI GPT-4o and GPT-4o mini GPT-4o and GPT-4o mini Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ By Value metric extraction for Quality AI GPT-4o and GPT-4o mini GPT-4o and GPT-4o mini Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ By Hold Adherence GPT-4o and GPT-4o mini GPT-4o and GPT-4o mini Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ By Transfer Adherence GPT-4o and GPT-4o mini GPT-4o and GPT-4o mini Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Churn & Escalation Identification GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Conversation Phase Identification GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Crutch Word Usage Detection GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Default Script Adherence GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Agent Answer Adherence and Customer Trigger Detection GPT-4o and GPT-4o mini GPT-4o and GPT-4o mini Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Generating Similar QM Utterance Suggestions GPT-3.5 Turbo GPT-3.5 Gemini 3, 2.5 Pro, Flash, Flash-Lite ❌ Post Conversation Sentiment Analysis GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Sentiment Analysis GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅ Topic Modelling GPT-4o GPT-4o Gemini 3, 2.5 Pro, Flash, Flash-Lite ✅
Gemini models support all Quality AI features. You must configure custom prompts for each feature. Default prompts are not supported.
Features Advanced Topic Discovery based on Custom Taxonomy and Resolution Detection Extracts topics and intents from agent conversations and detects resolution. Surfaces trends and anomalies based on customer sentiment, handle time, and resolution outcomes. Learn more .
Agent Empathy Identification Identifies instances of agent empathy in conversations where customers expressed negative sentiment. Learn more .
By Value Adherence Validation Validates data adherence by comparing extracted metric values against reference data using multilingual LLM analysis. Returns binary adherence scores per metric. Learn more .
Extracts by-value metrics from user, AI Agent, and human agent messages. Learn more .
By Hold Adherence Detects when an agent places a user on hold and resumes the conversation. Learn more .
By Transfer Adherence Identifies whether the agent informs the customer about a transfer. Learn more .
Churn & Escalation Identification Identifies customer churn risk and escalation intents across agent conversations. Learn more .
Conversation Phase Identification Identifies conversation phases for phase-level sentiment trajectory and resolution-aware sentiment analysis.
Crutch Word Usage Detection Detects agent crutch word usage in customer conversations. Learn more .
Default Script Adherence Determines agent adherence to default script steps: greeting, branding, privacy policy, customer verification, hold etiquette, and call closing. Learn more .
GenAI-based Agent Answer Adherence and Customer Trigger Detection Verifies agent adherence and detects customer triggers without requiring utterance configuration or training. Learn more .
Generating Similar QM Utterance Suggestions Generates semantically similar phrases during design time. Learn more .
Post Conversation Sentiment Analysis Generates post-interaction sentiment and emotion moment insights for agent conversations. Learn more .
Sentiment Analysis Displays a real-time sentiment indicator during active chat conversations, updating dynamically as the conversation progresses. Learn more .
Topic Modelling Extracts popular topics and intents from agent conversations. Learn more .
