Agent performance gives you a clear, structured view of how your AI Voice Agent (AIVA) is performing, so you can identify what is working, what needs improvement, and how to optimize outcomes over time. With this dashboard, you can monitor performance, uncover gaps, and take action without reviewing individual call recordings.
How to access Agent performance
Steps:
- Go to Aircall Dashboard > AI Agents > AI Voice Agents.
- Select your agent.
- Click Performance under the See how it's performing section.
Note: Agent Performance is currently in Beta and available to selected customers. If you do not see it in your sidebar, contact your Aircall Account manager.
What you will see on the dashboard
The Performance dashboard is organized into three main areas:
- Core performance metrics
- Productivity metrics
- Optimization suggestions
Each section provides actionable insights into your agent’s performance.
Core performance metrics
These metrics provide a quick overview of how effectively your agent is handling calls.
| Metric | Description |
|---|---|
| Autonomous resolution rate | The percentage of calls handled end to end by AIVA without human intervention. This reflects how much workload your agent removes from your team. |
| Call transfer rate | The percentage of calls transferred to a human agent, either as part of a configured workflow or due to fallback when the agent cannot resolve the request. A rising rate may indicate areas for improvement. |
| Goal completion rate | Measures whether the agent successfully completed the task it was designed to perform, regardless of whether further assistance was needed. |
| Knowledge gap highlight | Identifies the most impactful missing topic in your agent’s knowledge. It shows how many calls were affected, helping you prioritize improvements. |
Select See all suggestions to explore the full list in the Optimization suggestions section.
The call distribution diagram
This visual diagram shows how calls flow through AIVA and where they end. It helps identify patterns such as high escalation or unresolved rates.
| Category | What it means |
|---|---|
| Total calls | All calls received by AIVA during the selected time period |
| AI Agent only | Calls handled بالكامل by AIVA with no human involvement |
| Autonomous resolution | Calls resolved without live agent support |
| Handoff | Calls transferred to a human agent |
| Unresolved | Calls that ended without resolution due to dissatisfaction or unanswered requests |
Note: Call counts in this diagram are based on a calculated subset of your total calls. Totals may not exactly match overall call volume.
Productivity metrics
These metrics quantify the impact of your AI Voice Agent.
| Metric | Description |
|---|---|
| Total time saved | The total number of minutes your team saved because AIVA handled calls autonomously. This is useful for demonstrating return on investment. |
| Average call duration | The average length of calls handled by AIVA. Tracking this over time helps assess efficiency improvements. |
Optimization suggestions
This section provides prioritized, data-driven recommendations to improve your agent.
Knowledge gaps
Knowledge gaps are topics your callers frequently ask about that your agent cannot currently answer. Addressing these gaps reduces escalations and unresolved calls.
Each suggestion includes:
- A clear recommendation
- A Why it's important explanation with impact metrics
- Sample customer questions that triggered the gap
Steps to resolve a knowledge gap:
- Review the recommendation.
- Navigate to Teach it what to know in the left sidebar.
- Add the relevant content to your Knowledge Base.
- Save your changes so the agent can use the new information in future calls.
Tip: Start with the top suggestion in the list. Suggestions are ranked by impact, so addressing the first items will improve performance the most.
Configuration gaps
Configuration gaps are recommendations based on how your AI Voice Agent is set up and behaves during calls. Unlike knowledge gaps (which focus on what your agent knows), configuration gaps surface issues with how your agent is performing, such as response timing, greeting length, or follow-through on tasks.
Each suggestion includes:
- A clear recommendation with a suggested action
- A Why it's important explanation describing the observed behavior
- An estimated impact on your automated resolution rate
Steps:
- Review the recommendation and the behavior it describes.
- Navigate to your AI Voice Agent's configuration settings.
- Apply the suggested changes, for example, tightening prompt instructions or enabling retries for agent responses. For details on how to configure your AI Voice Agent, please see Configuring your AI Voice Agent.
- Save your changes and monitor performance in the Performance dashboard.
Tip: Each suggestion shows the approximate percentage of calls impacted and the potential improvement to your resolution rate. Use these to prioritize which gaps to fix first.
Providing feedback on suggestions
Each suggestion includes a thumbs up and thumbs down option. Use these to indicate whether the recommendation was helpful. Your feedback improves future suggestions.
Exporting your data
To share performance insights or build reports:
Steps:
- Go to the Performance dashboard.
- Click Export in the top right corner.
- Download your metrics for sharing or analysis.
Quick reference: metric definitions
| Metric | Definition |
|---|---|
| Total calls | Total number of calls received during the selected period |
| Autonomous resolution rate | Percentage of calls handled بالكامل without human intervention |
| Call transfer rate | Percentage of calls transferred to a human agent |
| Goal completion rate | Percentage of calls where the agent completed its assigned task |
| Total time saved | Total minutes saved through autonomous call handling |
| Average call duration | Average length of calls handled by AIVA |
| Knowledge gaps | Missing topics ranked by number of affected calls |
FAQs
What is the difference between Call transfer rate and Unresolved calls?
Call transfer rate includes all transfers to human agents, including intentional ones. Unresolved calls are those that ended without a satisfactory outcome.
Why does the call count in the diagram not match total calls?
The diagram reflects a calculated subset of calls, while the total calls metric includes all calls received.
How often are knowledge gap suggestions updated?
Suggestions update continuously as new calls are processed. Check regularly, especially after updating your Knowledge Base.
Can I share the Performance dashboard with my team?
Admin users with access to AIVA can view the dashboard. You can also export data to share with others.
Aircall Status
Help logging in