Metrics to measure whether your AI voice agent is working
Short answer: to measure an AI voice agent, track time to first contact, contact rate, qualification rate, meetings booked per 100 leads, show rate, conversion to sale or enrollment, cost per meeting and loss reasons. Always compare against what your team did before. ReshapeOS shows these metrics per campaign in its dashboard, with sentiment analysis of every conversation.
The 8 metrics
| Metric | How it's calculated | What it tells you |
|---|---|---|
| Time to first contact | From lead arrival to first call | Whether you get there before competitors |
| Contact rate | Leads reached / total leads | Whether call times and retries work |
| Qualification rate | Qualified leads / leads reached | Campaign quality |
| Meetings per 100 leads | Meetings booked / leads × 100 | The agent's main result |
| Show rate | Meetings held / meetings booked | Whether confirmations and reminders work |
| Conversion to sale or enrollment | Enrollments / meetings held | The business impact |
| Cost per meeting | Total cost / meetings booked | What each opportunity really costs |
| Loss reasons | Grouped from the conversations | Which objections need work |
How to read them
- Low contact rate: review call times, number of attempts and add WhatsApp.
- Low qualification rate: the problem is usually the campaign, not the agent.
- Few meetings despite good qualification: adjust how the agent proposes the appointment.
- Low show rate: strengthen confirmations and reminders.
Listen to conversations
The numbers tell you what's happening; the conversations tell you why. Listen to 20 calls a week at the start and adjust the script based on what you hear.
Frequently asked questions
What's the most important metric?
Meetings booked per 100 leads, followed by how many of those meetings turn into enrollments or sales.
How long does it take to evaluate an agent?
At least a week with real leads, and ideally a full month to compare against the previous one.
How do I compare it with my team?
Use the same metrics on leads from the same period the month before, or split one campaign between the agent and your team.
Try an agent