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

# Outcome

> Choose what Vocily retains and extracts after a conversation ends.

The **Outcome** tab controls what happens after a conversation finishes. It has three areas:
**Saved recordings and transcripts**, **AI Analysis**, and **Custom Analysis**.

## Saved recordings and transcripts

### Save Call Audio

Turn on **Save Call Audio** to keep a playable recording for review in [Conversation
History](/conversations/history). Use recordings to review the conversation, coach your team,
and investigate call quality.

### Transcript Storage

Turn on **Transcript Storage** to keep the conversation transcript available in Conversation
History for review, analysis, and follow-up. Both options are enabled by default for new agents.
Choose the settings that fit your team's review and privacy needs.

## AI Analysis

The built-in **AI Analysis** controls are:

* **Call Summary** — generates a concise English recap after the conversation.
* **User Sentiment** — classifies the caller's overall sentiment as `positive`, `neutral`, or
  `negative`.

The analysis also records a high-level outcome such as `interested`, `not_interested`,
`needs_followup`, `resolved`, or `unknown`, and identifies the primary conversation language.
Results appear in Conversation History after processing. Very short conversations may not have
enough information for meaningful analysis.

If analysis fails, the conversation remains available and the result is marked unavailable rather
than blocking finalization.

## Custom Analysis

Use **Custom Analysis** when the built-in summary and sentiment fields do not capture the
business outcome you need. Examples include qualification, objections, next steps, disposition,
or whether a customer provided a required detail.

### Organize analysis into categories

Create categories such as `Qualification`, `Objections`, or `Next steps`. Categories can be
enabled or paused independently.

### Add analysis items

Each item has a name, a prompt describing what to extract, and an answer format:

* **Free text** for an explanation or concise extracted value.
* **Predefined** for a controlled list of options.
* **Both** when the item needs a selected option plus supporting text.

Free-text items can use the expected formats **text**, **timestamp**, **numeric**, **boolean**, or
**email**. Predefined items can include option values and condition prompts. Keep each item
focused on one question so the result is consistent across conversations.

### Test before enabling

After adding at least one category and item, select **Test**, paste a representative transcript,
and inspect the structured result. This is the fastest way to find ambiguous prompts or options
that overlap.

Turn on **Custom Analysis** only after the test looks correct. When enabled, Vocily analyzes each
conversation using your categories and items and displays the structured result with the
conversation. Turning it off stops new analyses but keeps your categories and items so you can
revise and enable them again later.

## Use outcomes in workflows

Post-call custom APIs can run after custom analysis completes, while other APIs can run when the
conversation, recording, or built-in analysis completes. See
[Capabilities](/building-agents/capabilities) for workflow phases and trigger points.

## Save and review

Save the agent after changing the recording, transcript, or built-in analysis switches. Custom analysis
categories and items are created and edited from their own controls. Review the result in
[Conversation History](/conversations/history), then refine the prompts when the extracted
outcome is too vague, missing, or overconfident.