AI voice agent analytics
View performance metrics and analytics for your AI voice agent to track call handling effectiveness and guest satisfaction.
On this page
Before you begin
- Your organization must have HyperConnect enabled with AI voice agent provisioned.
Permission: Go to Configuration > Security > Security Roles > select a role > scroll to the Zenoti HyperConnect section. The following permissions are required:
- Connect Access Enable β Base access to HyperConnect.
- View Dashboard β Required to access analytics dashboards.
- Connect Reports Enable β Required to view reports.
- AI voice agent calls must be active to generate analytics data.
- Analytics events are collected asynchronously during calls and do not affect real-time conversation performance.
Audience: Center managers and operations administrators who monitor AI voice agent performance.
Steps
Access AI voice agent analytics
- Open HyperConnect from Chat in the Zenoti admin interface.
- Navigate to Analytics > AI Voice Agent Performance.
- Select your date range using the date picker controls.
- Optionally filter by specific call modes or agent performance criteria.
Review call performance metrics
The dashboard tracks four event types for each AI voice agent call: call initiation (when the AI voice agent answers), call acceptance (when the guest begins interacting), individual utterances with their classified intents, and call completion with the final outcome category.
- View the call completion dashboard showing:
- Total calls handled β Number of calls processed by the AI voice agent.
- Successful completions β Calls resolved without human handoff.
- Agent handovers β Calls transferred to human agents.
- Abandoned calls β Calls where guests disconnected during interaction.
- Review intent classification accuracy showing how well the system understood guest requests.
- Check average call duration and resolution times across different request types.
Note: Data is near real-time with a maximum delay of 60 minutes.
Analyze guest feedback data
Post-call feedback is collected from guests who interacted with the AI voice agent. When a guest completes a call, the system may send an SMS feedback prompt to the guest's phone number. Replies are correlated with the original call session.
- In the Feedback Analytics section, review:
- Total feedback count β Number of post-call feedback responses received.
- Average rating β Mean satisfaction score from guest feedback.
- Rating distribution β Breakdown of 1β5 star ratings.
- Will use again percentage β Guests who indicated they would use the AI voice agent again.
- Click on individual feedback entries to view detailed guest comments and ratings.
- Review feedback trends over time to identify improvement opportunities.
Note: The feedback SMS prompt is sent only when post-call feedback collection is enabled for your center. If the Feedback section shows no data, confirm with your administrator that feedback collection is configured.
Monitor conversation quality
- Access the Call Analysis section to view:
- Professionalism scores β AI-generated ratings for professional interaction quality.
- Empathy ratings β Assessment of the agent's empathetic responses.
- Efficiency metrics β How quickly the agent resolved guest requests.
- Communication effectiveness β Clarity and accuracy of AI responses.
- Review sentiment analysis showing positive, neutral, and negative interaction distribution.
- Examine specific call transcripts for detailed conversation analysis.
Tip: For Zeenie AI assistant analysis of call transcripts and agent performance, see the Zeenie category.
Track booking success rates
- In the Booking Performance section, monitor:
- Booking completion rate β Percentage of booking requests successfully processed.
- Rescheduling success β Rate of successful appointment changes.
- Cancellation handling β Effectiveness of cancellation request processing.
- Compare AI voice agent booking performance against human agent benchmarks.
- Identify common booking failure points for system optimization.
Review AI voice agent performance scores
- Navigate to the AI Voice Agent Performance section of the analytics dashboard.
- Select a date range and optionally filter by a specific call mode.
- Review agent performance scores across four dimensions:
- Professionalism β Rated on a scale of 0β10.
- Empathy β Rated on a scale of 0β10.
- Efficiency β Rated on a scale of 0β10.
- Communication β Rated on a scale of 0β10.
- Review individual feedback records, including guest ratings, free-text responses, and the associated call session references.
Note: Performance scores are derived from AI-powered call transcript analysis. Scores are available only for calls of sufficient duration. If a call is too brief, scores will show as unavailable.
Generate performance reports
- Select Export Analytics to download performance data.
- Choose your export format (CSV or PDF).
- Specify the date range and metrics to include in the export.
- Use exported data for management reporting or system optimization analysis.
Tips
- Review analytics weekly to identify trends in AI voice agent performance and guest satisfaction.
- Use feedback data to refine conversation flows and improve guest experience.
- Monitor handover rates to determine if additional AI training is needed for specific request types.
- Compare performance across different time periods to measure improvement over time.
- Filter the dashboard by call mode to compare performance across different interaction types such as booking, rescheduling, and cancellation.
- Use the Will use again percentage metric as a leading indicator of overall guest satisfaction with the AI voice agent experience.
Troubleshooting
No feedback data appears in the Feedback Analytics section. Post-call feedback SMS collection must be enabled for your center. If the section is empty, confirm with your administrator that feedback collection is configured. If feedback collection is not configured, the section will display a message indicating that feedback collection is not set up, with a link to the feedback configuration settings.
Call analysis scores show as unavailable. Performance scores require calls of sufficient duration to analyze. Very short calls β such as those where the guest disconnects immediately β do not generate scores. This is expected behavior and is not an error.
Analytics data appears delayed or incomplete. Analytics events are collected asynchronously and do not block real-time call handling. Data is near real-time with a maximum delay of 60 minutes. If data appears missing beyond this window, contact your administrator to verify that AI voice agent analytics collection is active for your center.