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

# Measure AI Visibility and Review Answer Evidence

> Configure prompts and research engines, run benchmarks, review citations and brand accuracy, and inspect answer readiness and AI referrals.

Open [AI Visibility](https://portal.wowagent.ai/seo-research/prompts) from **SEO Intelligence > AI Visibility** and select your website.

The current workflow includes **Overview**, **Prompt Library**, **Runs & History**, **Citations**, **Brand Accuracy**, **Answer Readiness**, and **AI Referrals**.

## Configure the prompt library

In **Prompt Library**, enter saved customer questions, one per line. You can leave the library empty to generate a benchmark from your website topics.

Set the language code, country code, prompts per run, repeats, and daily call limit. Repeats support values from one to three.

Choose available **WOWagent research engines**. Unavailable engines cannot be selected. An empty selection uses the available managed engines.

Choose a schedule: **Manual**, **Every 12 hours**, **Daily**, **Every two days**, or **Weekly**. Add competitor domains when relevant, review the alert setting, and click **Save settings**.

> [!NOTE]
>
> Scheduled research also depends on the automatic research control in [Crawlers and Jobs](/crawlers-and-jobs.md), engine availability, plan access, and credits.

## Run and compare results

Use **Run question** to investigate one question or **Run saved benchmark** to measure the saved set.

Research runs are queued and processed in the background. Review the status and completion time in **Runs & History**. Open a past run to inspect its answers.

Keep prompts, language, country, engines, and sampling settings consistent when comparing runs. An unavailable or failed execution is different from an answer that was successfully measured and did not mention your brand.

## Review citations

Use **Citations** to inspect the sources attached to measured answers. Independently retrieved web pages are supporting research, not proof that an engine cited them.

Read the actual answer and source before treating a mention as a recommendation.

## Review brand accuracy

In **Brand Accuracy**, save brand aliases and dated source facts using this format:

```text theme={null}
Our office is in Pune | https://example.com/contact | 2026-10-23
```

For an answer, choose a source fact, select **correct**, **incorrect**, or **unresolved**, and add supporting evidence. Click **Save evidence review**.

Saving a fact URL does not automatically verify its contents. Accuracy figures summarize reviewed claims; unknown or unreviewed claims are not automatically incorrect.

## Answer readiness and content briefs

Use **Answer Readiness** to inspect opportunities in your website content. Where available, create a content brief from a measured answer, then review it before using it in [AI Content](/ai-growth-and-content.md).

A readiness recommendation does not publish or edit your website automatically.

## AI referrals

Use **AI Referrals** to review available referral evidence. Missing analytics or unavailable measurements should be read as missing data, rather than zero visits.

For general keyword, competitor, and backlink research, see [SEO Intelligence](/seo-intelligence.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.