> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getthread.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Voice AI Summary Dashboard

> Report on Voice AI call volume, AI-answered versus rule-routed calls, ticket creation, talk time, caller satisfaction, top callers, call logs, and a day-and-hour heatmap.

The Voice AI Summary dashboard reports on your AI-powered phone system — the calls it answers, how they were routed, which became tickets, and how callers felt about them. Use it to size your phone channel, spot your highest-touch accounts, and see when call traffic actually happens.

Open it from **Thread admin → General → Analytics**.

## What this dashboard answers

* How many calls are we handling, and how many were **answered by AI** versus **routed by rule**?
* How many calls **turn into tickets**?
* Which **end customers** are generating the most call activity?
* What happened on a **specific call**?
* **When** does our call traffic arrive — and how much of it is after hours or at the weekend?

## Filters and time windows

| Window       | Where it's used                           |
| ------------ | ----------------------------------------- |
| Last 30 days | The Overview snapshot                     |
| Monthly      | Calls over time                           |
| Full history | Top customers, call logs, and the heatmap |

## What's on the dashboard

### Overview

A snapshot of the last 30 days: total call volume, the split between AI-answered and rule-routed calls, how many calls became support tickets, average call length, caller satisfaction ratings, and customer sentiment on the resulting requests.

### Voice AI calls over time

A monthly trend of call volume, broken out by inbound versus outbound and — for inbound — by how each call was handled: answered directly by AI, or routed by rule, such as after hours.

### Top customers by call volume

Ranks the end customers generating the most call activity, with inbound and outbound splits, how many of their calls became tickets, and average and total talk time. This is the fastest way to spot a high-touch account whose contract doesn't reflect its phone usage.

### Voice AI call logs

A searchable log of individual calls — who called, when, how long, which agent handled it, whether it created a ticket, the caller's satisfaction rating, and any linked ticket details including priority, status, and summary.

### Voice AI call heatmap

Call volume plotted by day of week against hour of day, so your real traffic pattern is visible — including after-hours and weekend concentrations that justify overflow coverage.

## Metric definitions

| Metric                    | Definition                                                                                                                  |
| ------------------------- | --------------------------------------------------------------------------------------------------------------------------- |
| AI-answered               | The call was handled directly by Voice AI.                                                                                  |
| Rule-routed               | The call was directed by a routing rule, such as an after-hours or overflow rule, rather than answered by the AI attendant. |
| Calls that became tickets | Calls that resulted in a support request being created in Thread.                                                           |
| Talk time                 | Call duration, reported as both an average and a total per customer.                                                        |
| Caller satisfaction       | The rating the caller gave when the feedback step ran on that call type.                                                    |
| Sentiment                 | The tone of the resulting support conversation.                                                                             |

## Things to watch

<Note>
  **Not every inbound call creates a ticket, and that's expected.** Plenty of calls are answered, resolved, or handed off without a support request being raised. A gap between call volume and ticket volume is normal behavior, not missing data.
</Note>

* **A missing satisfaction rating means the caller wasn't asked** — the feedback step only runs where it's enabled for that call type. It does not mean the caller was unhappy.
* **Sentiment reflects the support conversation, not the call audio.** It's scored on the resulting request, so a difficult call with a smooth follow-up can read positive.
* **Rule-routed volume is a configuration outcome.** A large rule-routed share usually reflects your after-hours and overflow rules working, not the AI declining calls.
* **Data refreshes daily**, so a call from this morning generally appears tomorrow.

## Next steps

<CardGroup cols={2}>
  <Card title="Voice AI attendant modes" icon="phone" href="/ai-agents/voice-ai-attendant-modes">
    Understand what drives the AI-answered versus rule-routed split.
  </Card>

  <Card title="Overflow and after hours" icon="moon" href="/ai-agents/voice-ai-setup-phase-2-overflow-after-hours">
    Act on what the heatmap shows about out-of-hours traffic.
  </Card>
</CardGroup>


## Related topics

- [ROI Value Realization Dashboard](/analytics/roi-value-realization-dashboard.md)
- [Triage Agent Summary Dashboard](/analytics/triage-agent-dashboard.md)
- [Voice AI FAQ](/ai-agents/voice-ai-faq.md)