> For clean Markdown of any page, append `.md` to the page URL.
> For a complete documentation index, see https://docs.sarvam.ai/llms.txt.
> For full documentation content in one file, see https://docs.sarvam.ai/llms-full.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.sarvam.ai/_mcp/server.

# Goals

Goals is a tab under **Monitor → Agent Analytics**. It shows whether your agents are meeting their configured goals, and compares the calls that met the goal against those that did not — so you can see which call patterns correlate with success.

Open **Monitor → Agent Analytics** and select the **Goals** tab.

The Goals tab appears only when the selected agent has a [goal](/conversations/build/process#goals) configured, and is available for voice agents. Each agent can have one goal; if none of the selected agents have a goal configured, the tab prompts you to pick one.

## What counts as a goal

| Term                      | Meaning                                                                                 |
| ------------------------- | --------------------------------------------------------------------------------------- |
| **Goal-eligible call**    | A connected call that was evaluated against the goal (passed, failed, or not evaluated) |
| **Goal achieved**         | A goal-eligible call where the goal was met (passed)                                    |
| **Goal achievement rate** | Goal achieved ÷ goal-eligible, as a percentage                                          |

## KPI tiles

| Tile                      | Meaning                                                                                                        |
| ------------------------- | -------------------------------------------------------------------------------------------------------------- |
| **Goal achievement rate** | Share of goal-eligible calls where the goal was met, with the achieved-of-eligible counts alongside            |
| **Avg turns to goal**     | Average number of conversation turns on goal-achieved calls, with the not-achieved cohort shown for comparison |

## Achieved vs not achieved

Two side-by-side cohorts — **Goal achieved** and **Goal not achieved** — compare the calls that met the goal against those that didn't. Each cohort shows its call count and share of goal-eligible calls, plus:

| Metric                 | Meaning                                      |
| ---------------------- | -------------------------------------------- |
| **Avg call duration**  | Average length of a call in the cohort       |
| **Avg turns per call** | Average conversation turns in the cohort     |
| **Avg latency**        | Average agent response latency in the cohort |

The **Goal not achieved** cohort also shows how far each metric differs from the achieved cohort, so you can see whether failed calls run longer, take more turns, or respond more slowly.

## Breakdowns

| Chart                               | What it shows                                                                       |
| ----------------------------------- | ----------------------------------------------------------------------------------- |
| **Turn distribution by cohort**     | How achieved and not-achieved calls spread across turn buckets (1, 2, 3–4, 5–8, 9+) |
| **Language split by cohort**        | Achievement compared across the top languages                                       |
| **Goal achievement by time of day** | Achievement rate by hour of day (IST), with the peak and trough hours called out    |
| **TTS voice × goal rate**           | Achievement rate per speaker voice, to see which voice performs best                |
| **Intro message × goal rate**       | Achievement rate per opening line, ranked, to see which intro converts best         |

## How to use Goals

#### Start with the achievement rate

Read the **Goal achievement rate** for the period. Use the achieved-of-eligible count to judge how much data it is based on.

#### Compare the two cohorts

Check whether not-achieved calls run longer, take more turns, or have higher latency than achieved calls.

#### Look for patterns

Use the turn, language, time-of-day, voice, and intro breakdowns to find what correlates with success.

#### Confirm on real calls

Open [Call Logs](/conversations/monitor/agent-analytics/call-logs) with the same filters and review a few not-achieved calls before changing the agent.

## Where to go next

#### [Engagement](/conversations/monitor/agent-analytics/engagement)

Check whether connected calls become real conversations.

#### [Call Logs](/conversations/monitor/agent-analytics/call-logs)

Open the individual calls behind the goal numbers.

#### [Overview](/conversations/monitor/agent-analytics/overview)

Return to the period-level health check.