# AI visibility scans

> Run an AI visibility scan, understand the visibility score and share of voice, and see which prompts and engines actually cite your Shopify store.

Canonical URL: https://sphen.ai/help/ai-visibility/
Last updated: 2026-08-14

## What it does

An AI visibility scan asks real buyer questions to the AI engines your customers use and records whether your store appears in the answer, where it ranks, and which other domains got cited instead. One full scan covers five buyer prompts across five AI engines. Results land on three pages of their own — **Rankings**, **Prompts** and **Sources** — which have no navigation entry: you open them from the **AI visibility deep dives** dropdown in the tab row on **Analytics**.

### Why it matters

Rank tracking tells you where you sit in a list of blue links. It says nothing about whether ChatGPT names your store when a buyer asks for a recommendation. The scan turns that into a number you can act on: if competitors and review sites are cited on your core prompts and you are not, the fix is a content and citation problem, not a bidding problem.

## Setup

1. On **Overview**, find the **AI visibility scan** card and press **Start AI Visibility scan**. The scan runs in the background — you can leave the page.
2. Review which questions are being asked. On **Analytics**, open **AI visibility deep dives → Prompts** and use the **Custom prompts** tab to see the prompt panel Sphen generated from your catalog.
3. Add the questions your buyers really ask under **Add a custom prompt**: enter the prompt, pick an **Intent** and a **Language**, then press **Add prompt**. It joins the next scan.
4. Delete prompts that don't reflect your market with **Delete**. Results already collected stay in past runs.
5. When the scan finishes, open **Analytics** and work through **Prompts**, **Rankings** and **Sources** in that order — each from the **AI visibility deep dives** dropdown.
6. Use **Sources** to build an outreach list: the domains AI answers cite instead of you. Send them to the workflow described under [Mentions](/help/mentions).

## Reading the results

The **Visibility score** is a sampled average: the share of sampled AI answers across engines that mention your store, scored out of 100. It is not an exhaustive crawl, so read the trend across scans rather than a single number. **Share of voice** answers a narrower question — of the store links AI answers cited, how many pointed at you.

**Win rate** is the sharpest metric on the page: of the answers that do mention you, the share that recommend you as the top pick. Being mentioned but never winning is a positioning problem, not a visibility problem.

Under **Analytics → Browser checks** you see the same questions asked in real guest-browser sessions of the ChatGPT and Gemini web interfaces, as a reality check against the API samples. In **Rankings**, empty cells in the engine-coverage matrix are gaps worth working. A healthy picture: your own domain climbing the **Cited domains** table while the review sites above you turn into your outreach targets.

## Troubleshooting

### The scan has been queued for a long time.

The card shows **Scan is taking longer than expected** when a run is queued behind others. It finishes in the background; press **Refresh** or check **Activity** for the job. See [Troubleshooting](/help/troubleshooting).

### I hit my monthly scan limit.

Each plan includes a fixed number of full scans per month and the allowance resets at the start of the next month. Check what is left under **Settings → Plans**; see [Plans & billing](/help/billing).

### No results at all after a completed scan.

That is a real result, not an error: your store was not mentioned in any sampled answer. Start with [Recommendations](/help/recommendations) to close catalog gaps, then [Content](/help/content) and [Mentions](/help/mentions) to build citable material.

## Frequently asked questions

### Which engines are checked?

Every plan includes five AI engines. Per-engine results are on the **Rankings** deep dive under **Engine coverage**, and the guest-browser checks cover the ChatGPT and Gemini web interfaces.

### Why does the score move between scans without me changing anything?

AI answers are non-deterministic and the scan is a sample, not a census. Small movements are noise. Look for direction across several scans, and read **Gainers** and **Losers** on the **Prompts** deep dive for week-over-week movement that survived sampling. The **AI Shopping** tab carries the same movers per product, where that tab has been switched on for your store.

### Can I track a competitor?

Indirectly, and that is usually enough. Competitors that appear in the same answers are listed per prompt under **Competitors mentioned** in the prompt details, and competing domains show up in **Cited domains**.

## Still stuck?

Support answers installation, catalog sync, audit, billing, privacy, and account questions for the SPHEN.AI Shopify app. https://sphen.ai/support/