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

# The seven AI engines Are you found by AI? measures

> The full measurement covers ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek and Google AI Overviews. The free scan asks ChatGPT and Gemini. See how each engine differs and what counts as being named.

Your customers are not all using the same AI assistant. Some open ChatGPT, some read Google's AI panel, some rely on Perplexity for research, and others use Claude, Grok or DeepSeek. A business that appears on one engine but not the others has fragile visibility. The full Are you found by AI? measurement asks your buyer questions across seven engines, saves every answer, and shows which engines named you, who they named instead and the words behind each result.

## The seven engines

| Engine | Behaviour and sourcing notes |
| - | - |
| **ChatGPT** (OpenAI) | Can search the web before answering buyer questions. Where it returns its search queries, we keep them with the answer. |
| **Perplexity** | Retrieval-first: runs live web searches before answering and returns the sources it used. Sensitive to fresh, indexed content. |
| **Gemini** (Google) | In our measurement, Gemini uses its own knowledge without Google Search grounding, so its answers reflect what the model already knows about your business. |
| **Claude** (Anthropic) | Balances model knowledge with retrieval. Favours well-attributed, credentialled content. |
| **Grok** (xAI) | Draws on current web content and conversations on X. |
| **DeepSeek** | Favours structured, citable passages from authoritative domains. |
| **Google AI Overviews** | Google's AI-generated answer panel above organic results. It issues related searches before composing an answer, and typically names a small number of businesses or resources. |

<Note>
  The **free scan** asks two engines, ChatGPT and Gemini. **Pro** asks your buyer questions across all seven engines. Available answers and sources depend on the completed measurement, and engine availability varies.
</Note>

<Tip>
  Search access and training preferences are separate controls in `robots.txt`. Review the search permissions you intend (for example `OAI-SearchBot`, `Claude-SearchBot` and `PerplexityBot`) separately from training preferences. The [free AI crawler checker](https://areyoufoundbyai.com/tools/ai-crawler-checker) shows both.
</Tip>

## Why engines disagree

AI engines are not a single oracle. The same buyer question asked across all seven on the same day can produce seven different shortlists, and one engine asked twice can produce two different answers. Three factors drive the difference:

**Answers vary between runs.** An engine may be asked the same question more than once, and every saved answer counts on its own. There is no majority vote. Repeating the same buyer questions over time helps you tell an isolated answer from a wider trend.

**Source differences.** Each engine answers from a different mix of its own model knowledge, the live web and platform-specific data. A business with fresh press coverage may appear on a retrieval-first engine such as Perplexity before it appears on an engine answering from model knowledge alone.

**Freshness windows.** Engines that search the web can surface new mentions within days. Model knowledge without live retrieval updates on training cycles. A new listing or press mention can therefore change answers on some engines well before others.

## What counts as being named

The headline measure is the **named share**: countable saved answers that name your business, divided by all countable saved answers. A link to your website on its own does not count as your business being named in the answer. Missing or incomplete evidence is left out, never counted as a no.

Reports distinguish several outcomes, which can overlap:

* **Named:** your business is named in the answer.
* **Cited:** your website is linked in the answer's sources.
* **Listed:** your name appears in a list, heading or source title rather than in the prose.
* **Found via search:** a page appears in recorded search results. That alone is not a citation or a recommendation.

The older **AI Visibility score** (out of 100) uses a different basis: being named in the answer prose counts in full toward the visibility score, and a positive result recorded only as a list entry, citation or source title counts half. That score is a weighted diagnostic, not a percentage of AI recommendations, and it cannot be converted into a named share.

Every answer is saved in full with its model and date, and with the search queries the engine returned where it returned them.

<Tip>
  Open your report and read the saved answers for each engine. The sentences the engine actually gave tell you more than any single number: which questions lead to a recommendation, which competitors appear in the same answers, and whether you are named in the prose or only linked in the sources.
</Tip>


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