# Get Audience

> Fetch one audience's column structure and an optional preview of its rows. Read-only and free.

**Date:** 2026-09-21  
**Source:** https://gtm.ai/docs/mcp/tools/get-audience

---

## Overview

Look inside an audience before you act on it. Get Audience returns the full shape of one audience: every column, what type it holds, where it came from, and what you're allowed to do with it. Turn on the row preview and you see the actual data too, including which cells are still enriching. Inspecting first means downstream work is grounded in what the audience really contains.

## What You'll Get

- Audience metadata: name, type, origin, record count, and folder
- Column definitions: id, name, data type, source, and whether each column is filterable, sortable, editable, or exportable
- An optional row preview, with each cell's value and its state

## Best For

- Checking an audience's columns before building on it
- Confirming an enrichment run finished before exporting
- Spot-checking real values rather than trusting a record count
- Understanding how an audience someone else built is structured

You need an audience `id` to call this. Get one from [Browse Audiences](/docs/mcp/tools/browse-audiences).

## How It Works

Row preview is off by default. Turn it on to see up to 25 rows, with 5 returned unless you ask for more. Scope the preview to specific columns when an audience is wide.

Every previewed cell reports a state, so a real value is never confused with one that is still processing:

- `RESULT` holds valid data
- `BLANK` was never populated
- `LOADING` is still enriching
- `NO_RESULT` ran and found nothing
- `ERROR` failed for that cell

When a lot of cells read `LOADING` or `ERROR`, the audience is still processing. The model should say so rather than treat partial data as final.

## Technical Details

### Credits

Free. Get Audience is read-only and consumes no credits. See [Credits & Billing](/docs/mcp/credits).

### Requirements

The audience `id` comes from [Browse Audiences](/docs/mcp/tools/browse-audiences), and the model uses it verbatim rather than guessing.

### Automatic Tool Selection

The AI reaches for Get Audience after [Browse Audiences](/docs/mcp/tools/browse-audiences), once it knows which audience you mean. It reads the column schema to learn how the audience is built, then previews a few rows to see real data before planning anything downstream. The `rowId` and `columnId` values it returns are what [Manage Audience Rows](/docs/mcp/tools/manage-audience-rows) and [Manage Audience Columns](/docs/mcp/tools/manage-audience-columns) need for updates. No ID gets guessed. For analytical questions over the whole audience, such as breakdowns, rankings, or fill rates, the AI uses [Audience Analysis](/docs/mcp/tools/audience-analysis) instead of previewing rows.

## Example Prompts

> *"What columns are in my Dreamforce audience?"*

> *"Show me a few rows from that audience."*

> *"Preview the email and title columns for my SaaStr list."*

> *"Did the enrichment on this audience finish?"*