Overview

The AI Agent Insights Dashboard provides a centralized view of how your AI Agent is performing across customer interactions.

It helps you:

• Monitor conversation performance
• Understand user behavior and intent
• Identify gaps in agent responses
• Take action to improve outcomes

The dashboard is divided into four main tabs:

• Overview
• Engagement
• User Behavior
• Recommendations

Each tab surfaces specific insights through structured widgets.

Overview Tab

Business Overview

Executive Summary

This widget provides a high-level, AI-generated summary of your AI Agent’s performance for the selected time period.

Data Provided:

• Overall resolution rate
• Fallback rate (instances where AI could not resolve queries)
• Total number of conversations
• Topic-level performance highlights
• Key problem areas (e.g., high fallback topics, negative sentiment)
• Suggested improvement areas

How to Use:

• Quickly understand overall performance without analyzing raw data
• Identify priority areas needing attention
• Guide optimization strategy

AI Recommendations

Displays a summary of actionable recommendations identified for the selected period.

Data Provided:

• Number of recommendations requiring attention
• Quick access to detailed recommendations

How to Use:

• Determine if immediate action is required
• Navigate to the Recommendations tab for deeper analysis

Agent Conversation Insights

This section provides quick entry points into deeper analytics:

Topic Analysis

• Shows what users are asking about
• Helps identify high-frequency topics

Sentiment Trends

• Tracks how users feel about interactions over time

Volume & Usage

• Tracks total conversations processed

Engagement Tab

Engagement Overview

Total Conversations

Definition:

Total number of conversations handled by the AI Agent during the selected time period.

Data Provided:

• Absolute number of conversations
• Percentage change compared to the previous period

Use Case:

• Measure adoption and usage growth

Avg. Conversations per Day

Definition: Average number of conversations handled daily.

Use Case:

• Understand daily interaction load
• Identify consistency or spikes in usage

AI Resolution Rate

Definition: Percentage of conversations successfully resolved by the AI Agent without fallback.

Use Case:

• Measure the effectiveness of the AI Agent
• Track improvements over time

Conversation Trends

Definition: A time-series graph showing conversation volume over the selected period.

Data Provided:

• Daily conversation count
• Trends and spikes

Use Case:

• Identify peak usage periods
• Correlate spikes with campaigns or events

Device Breakdown

Definition: Distribution of conversations across device types.

Data Provided:

• Mobile vs Desktop usage percentages

Use Case:

• Optimize experience based on device preference

User Geography

Definition: Geographic distribution of users interacting with the AI Agent.

Data Provided:

• Conversations by region/country

Use Case:

• Understand where users are engaging from
• Support localization strategies

User Behavior Tab

User Sentiment Analysis

Definition: Breakdown of user sentiment across conversations.

Categories:

• Positive
• Neutral
• Negative

Data Provided:

• Percentage of conversations per sentiment
• Number of conversations
• Top drivers (topics influencing sentiment)

Use Case:

• Identify experience quality
• Detect dissatisfaction or friction points

Topic Analysis

Definition: Categorizes conversations into topics to understand user intent.

Data Provided per Topic:

• Number of conversations
• Percentage of total conversations
• Sentiment classification
• Resolution rate

Example Topics:

• Booking
• Cancellation
• Pricing
• Technical Issues

Use Case:

• Identify high-demand topics
• Understand performance per topic
• Detect gaps in knowledge

Add Topic

Users can define custom topics to improve classification.

Use Case:

• Align analytics with business-specific categories
• Improve insight relevance

Recommendations Tab

AI Recommendations

This tab surfaces actionable insights based on conversation analysis.

Recommendation States

Needs Attention

• Gaps where AI Agent could not resolve queries
• Requires user action

Resolved

• Issues that have been addressed

Dismissed

• Items marked as not relevant or duplicate

Data Provided

• Recommendation category (e.g., missing knowledge)
• Severity level
• Supporting conversation snippets
• Status (Needs Attention, Resolved, Dismissed)
• User actions taken

Filters

• Severity Filter: Prioritize critical issues
• Date Filter: Analyze recommendations by time period

Use Case

• Identify missing knowledge areas
• Improve AI response quality
• Track the resolution of issues over time

Best Practices

• Start with Executive Summary for quick insights
• Use Engagement tab to monitor performance trends
• Analyze User Behavior to understand intent and sentiment
• Regularly review Recommendations to improve the agent
• Update topics to ensure accurate classification