> For the full page index, fetch https://echonova.ai/llms.txt

# The agent overview tab

Every card, number and header control on an agent's Overview tab, and where each one comes from.

Overview is the tab an agent opens on. Nothing on it is editable: read it to see
whether the agent is live, how much work it did in the last 30 days, what it has
cost this month, and what it said in its last three conversations. Everything
you can change lives on the other tabs.

## The header

The header sits above every tab, so it is the same on **Persona**, **Voice** and
the rest.

| Control              | What it shows                                                                                                            |
| -------------------- | ------------------------------------------------------------------------------------------------------------------------ |
| Name and description | The agent's name and description. With no description it reads **No description**                                        |
| **Last updated:**    | When the agent was last changed                                                                                          |
| Status chip          | **Live** for a live agent, **Archived** for an archived one                                                              |
| **Pause Agent**      | Archives the agent. On an archived agent the same control reads **Restore**                                              |
| **Test**             | Opens the live playground. Disabled while the agent is archived, and reads **Starting…** while a session is being set up |

**Pause Agent** and archiving are the same action, and an agent is either live
or archived. An archived agent also gets a banner under the header saying it
cannot take calls or start a session until you restore it.

The tab row underneath carries **Overview**, **Persona**, **Voice**,
**Knowledge**, **Tools**, **Channels**, **Conversations** and **Settings**.

## The recommendation strip

A row of suggestion cards at the top of the tab. Each one is a shortcut to a
tab, not a report on how your agent is set up:

- **Get your agent ready**, shown while the agent has had no voice calls and no
  text conversations. **Continue setup** opens the **Persona** tab.
- **Add test coverage**, always shown. **Add tests** opens the live playground.
- **Connect a channel**, always shown. **Start Connect** opens the **Channels**
  tab.

For what the agent has actually done, read **Performance overview** below.

## Performance overview

Five cards. Until the agent has had at least one voice call or one text
conversation they are replaced by
`Performance data will appear after your agent goes live`, and the heading reads
**Overview numbers** rather than **Performance overview**.

| Card                   | Its detail line  | What it counts                                                                                                                          |
| ---------------------- | ---------------- | --------------------------------------------------------------------------------------------------------------------------------------- |
| **Voice calls**        | **last 30 days** | Voice conversations in the window                                                                                                       |
| **Text conversations** | **last 30 days** | Text conversations in the window                                                                                                        |
| **Unique customers**   | **last 30 days** | Distinct contacts across those conversations                                                                                            |
| **Avg. call duration** | **last 30 days** | Total voice seconds divided by the voice calls that reached an end, shown as seconds under a minute and as minutes and seconds above it |
| **Credits used**       | **this month**   | Credits charged since the start of the calendar month                                                                                   |

The first four read a rolling 30-day window. **Credits used** is measured from
the first of the current calendar month instead, which is why its detail line
says **this month**. Compare it against your plan, not against the other four
cards.

## Last test results

**Last test results** shows **No tests run yet** and a button that opens the
live playground. Test an agent by talking to it there: see
[Test your agent](/docs/agents/test-your-agent).

## Recent conversations

The three most recent conversations for this agent. Each row shows the customer,
the last message if there is one, the channel it came in on, the date and time
it started, and its status, with a failed status marked in red. A phone icon
means a voice conversation and a speech-bubble icon means a text one. Selecting
a row opens that conversation on the **Conversations** tab. **View All** goes to
the same tab without opening anything.

With no conversations yet, the section says
`Conversations will appear here after you connect a channel and your agent goes
live` and offers **Connect channel**.

## Worked example

An agent shows **Voice calls** 42, **Text conversations** 6,
**Unique customers** 31, **Avg. call duration** 1m 12s and **Credits used** 380.

That is 48 conversations from 31 people in 30 days, so callers are coming back.
An average call of 1m 12s on a support agent is short enough to suspect callers
are hanging up early, so open **Recent conversations**, read three transcripts
end to end, and check whether the agent is ending conversations before the
question is answered. The 380 credits are this month only; a 30-day comparison
lives on [Analytics](/docs/agents/analytics).

## When the numbers fail to load

The cards and the recent conversations are fetched together, so one failure
blanks both. The tab is replaced by **Couldn't load analytics** with the body
**Check your connection and try again.** and a **Retry** button that fetches
them again.

## FAQ

### Can I change the 30-day window?

Not on this tab. For other windows and a breakdown per channel, use
[Analytics](/docs/agents/analytics).

### Why does Credits used not match the other cards?

It counts the current calendar month, while the other four count the last 30
days. On the 2nd of the month it is almost always the smaller number.

### Why is Avg. call duration empty when I have had calls?

It only counts voice calls that reached an end. A call still running, or one
that failed before it ended, is not in the average.

### Does the Overview tab show text and voice together?

The counts are split into **Voice calls** and **Text conversations**.
**Unique customers**, **Credits used** and **Recent conversations** cover both.

### Nothing on this tab is editable. Where do I change things?

On the other tabs: **Persona** for wording, **Voice** for the voice and the
model, **Knowledge** for sources, **Tools**, **Channels** and **Settings** for
the rest.

## Related

- [What is an agent](/docs/agents/what-is-an-agent): what each tab holds
- [Create an agent](/docs/agents/create-an-agent): make one before you read its numbers
- [Write the persona](/docs/agents/persona): the first tab to edit
- [Analytics](/docs/agents/analytics): longer windows and a breakdown per channel
- [Test your agent](/docs/agents/test-your-agent): talk to it in the playground