Generative AI Features - Quality AI¶
Quality AI uses generative AI models to analyze agent–customer conversations in real time and post-conversation. It extracts insights, detects behavioral patterns, and evaluates agent performance to improve productivity, strengthen intent detection, support sentiment analysis, and deliver seamless end-user experiences.
Model Feature Matrix¶
The following table displays the features and the supported models.
(✅ Supported | ❌ Not supported)
| Feature | Azure OpenAI Model(s) | OpenAI Model(s) | Gemini Model(s) | Custom LLM |
|---|---|---|---|---|
| Advanced Topic Discovery based on Custom Taxonomy and Resolution Detection | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Agent Empathy Identification | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| By Value Adherence validation for Quality AI | GPT-4o, GPT-4o mini | GPT-4o, GPT-4o mini | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| By Value metric extraction for Quality AI | GPT-4o, GPT-4o mini | GPT-4o, GPT-4o mini | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| By Hold Adherence | GPT-4o, GPT-4o mini | GPT-4o, GPT-4o mini | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| By Transfer Adherence | GPT-4o, GPT-4o mini | GPT-4o, GPT-4o mini | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Churn & Escalation Identification | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Conversation Phase Identification | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Crutch Word Usage Detection | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Default Script Adherence | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| GenAI-based agent answer adherence and customer trigger detection | GPT-4o, GPT-4o mini | GPT-4o, GPT-4o mini | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Generating Similar QM Utterance Suggestions | GPT-3.5 Turbo | GPT-3.5 | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ❌ |
| Post Conversation Sentiment Analysis | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Sentiment Analysis | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
| Topic Modelling | GPT-4o | GPT-4o | Gemini 2.5 (Pro/Flash/Flash-Lite), Gemini 3.1 Pro Preview, Gemini 3 Flash Preview | ✅ |
Note
Gemini models support all Quality AI features. You must configure custom prompts for each feature. Default prompts are not supported.
Advanced Topic Discovery based on Custom Taxonomy and Resolution Detection¶
Extract topics and intents from agent conversations detect resolution of agent conversations to discover levers based on customer sentiment, Handle Time and Resolution and quickly surface trends and anomalies based on these outcomes. Learn more.
Agent Empathy Identification¶
Identify instances of agent empathy where customers have expressed negative sentiment through Quality AI using LLMs. Learn more.
By Value Adherence validation for Quality AI¶
Validates data adherence by comparing extracted metric values against reference data using multilingual LLM analysis. Performs precise value matching and returns binary adherence scores for each metric. Learn more.
By Value metric extraction for Quality AI¶
Extract by value metric from user/AI Agent/human agent messages through Quality AI using LLMs. Learn more.
By Hold Adherence¶
This feature detects when an agent places a user on hold and resumes the conversation. Learn more.
By Transfer Adherence¶
This feature identifies whether the agent informs the customer about the transfer. Learn more.
Churn & Escalation Identification¶
Utilize LLMs to identify agent crutch word usage in customer conversations using Quality AI. Learn more.
Conversation Phase Identification¶
Post conversation phase identification for phase level sentiment trajectory and resolution aware sentiment analysis.
Crutch Word Usage Detection¶
Use LLMs to detect agent crutch word usage in customer conversations through Quality AI. Learn more.
Default Script Adherence¶
Determine agent adherence to default script steps across greeting, branding, privacy policy, customer verification, hold etiquette, and call closing through Quality AI's conversation Intelligence dashboard using LLMs. Learn more.
GenAI-based agent answer adherence and customer trigger detection¶
Utilize LLMs to verify agent adherence and detect customer triggers without requiring configuration of utterances or training. Learn more.
Generating Similar QM Utterance Suggestions¶
Utilize LLM to generate similar phrases during the design phase, focusing on semantic similarity. Learn more.
Post Conversation Sentiment Analysis¶
Generate post-interaction sentiment and emotion moment insights for agent conversations from Quality AI using LLMs. Learn more.
Sentiment Analysis¶
Agents monitor customer sentiment dynamically during active chat conversations. When enabled, this feature displays a sentiment indicator that updates in real-time based on ongoing analysis of the conversation. Learn more.
Topic Modelling¶
Extract popular Topics and Intents that customers discuss across agent conversations through Quality AI using LLMs. Learn more.