# What Is an Answer Engine and Where Does Your Store Fit?

> What an answer engine is, the three ways it answers (knowledge base, snippet or generative answer) and where your store can show up.

URL: https://clione.ai/glossary/answer-engine/
Lang: en

"Answer engine" names a class of system, not a brand: any search engine that, instead of handing you ten links, **hands you the answer**. It is not a customer service chatbot or a synonym for ChatGPT, although ChatGPT with search is one. And it is not the same as [AEO (Answer Engine Optimization)](/glossary/aeo/), which is the practice of optimizing content for these systems. A store should care for a simple reason: every shopper question that gets resolved inside the engine is a visit that never arrives, unless the answer comes from your page and cites it.

## How it works

What defines an answer engine is where it gets the answer. There are three families, and a store has a different role in each.

| Family            | Where the answer comes from                    | Examples                                                                                                           | Can it cite your page?                               |
| ----------------- | ---------------------------------------------- | ------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------- |
| Knowledge base    | The system's own data, looked up or calculated | Google's [Knowledge Graph](/glossary/knowledge-graph/), Wolfram\|Alpha                                             | Almost never: the answer is not a piece of your site |
| Page extract      | A piece of a single page                       | Google's featured snippet                                                                                          | Yes, as the only source                              |
| Generative answer | Text composed from several retrieved pages     | [AI Overviews](/glossary/ai-overviews/), [AI Mode](/glossary/google-ai-mode/), [Perplexity](/glossary/perplexity/) | Yes, among several sources                           |

### Answering from a knowledge base

The first family does not search the web at the moment it answers. When Google introduced the Knowledge Graph, in May 2012, it described it as a model of things in the world and their relationships, with more than 500 million objects and more than 3.5 billion facts, which lets it **summarize the key facts about a topic** next to the results: for Marie Curie, say, when she was born and died, her education and her discoveries.[^google-knowledge-graph] Wolfram|Alpha takes the idea further: it says it gets answers not by searching the web but by doing computations over a large collection of its own data, algorithms and methods, in order to be a single source of definitive answers to factual questions.[^wolfram-about]

For a store, this family has a clear consequence: **a calculated answer comes from none of your pages**. If your customer's question is "how many grams are 8 ounces," none of your content is going to be the source.

### Extracting a snippet

The second family takes a piece of a page and puts it on top, with its link. It is the model of the featured snippet and of voice assistants, and the AEO entry explains it in detail. Here the page wins or loses whole: there is a single source.

### Composing a generative answer

The third family is the one that made the term popular, and it is the heart of [AI search](/glossary/ai-search/). The paper that named [GEO (Generative Engine Optimization)](/glossary/geo/) describes these systems, which it calls [generative engines](/glossary/generative-engine/), as a set of generative models plus a search engine that retrieves documents: they receive the query and, in several steps, generate a final answer **backed by the retrieved sources, with attributions inside the text**.[^geo-paper] Google describes its two features like this: AI Overviews help you grasp the essentials of a complicated topic and offer links to keep going, and AI Mode is mostly for queries that ask to explore, reason or compare options.[^google-ai-features]

## What changes for whoever publishes

In a classic search engine, the result is your page and the reader decides whether to click through. In an answer engine, the result is the answer, and your page is, at most, its source. The authors of the GEO paper framed it as a problem for content creators: with engines that work as black boxes and change fast, they **have little control over when and how their content is shown**.[^geo-paper] That is where the fear of [zero-click](/glossary/zero-click-search/) comes from, the search that ends without visiting any site.

What is in your hands, at least on Google, is being able to be a source. To appear as a supporting link in AI Overviews or AI Mode, a page has to be indexed and eligible to be shown in Search with a snippet, and there are no additional technical requirements; even so, meeting them does not guarantee that Google will crawl, index or show it.[^google-ai-features] Google adds that you do not need to create files for AI or special markup to appear.[^google-ai-features]

## In a store

A **fictional example**: three questions that reach a coffee store, and which family of engine usually answers them.

| Shopper's question                                             | Type of question    | Family that answers it       | Where does the store compete?                                                                    |
| -------------------------------------------------------------- | ------------------- | ---------------------------- | ------------------------------------------------------------------------------------------------ |
| How many grams are 8 ounces?                                   | Fact or calculation | Knowledge base               | Nowhere: the answer has no source                                                                |
| What is the difference between a light roast and a dark roast? | Definition          | Extract or generative answer | With a guide of its own that explains it better than the others                                  |
| Which grinder should I buy for espresso for under $100?        | Recommendation      | Generative answer            | With [product pages](/glossary/product-page/) that say what the shopper asks, as one more source | |

The third row is the one that sells, and it is exactly where a [generative engine](/glossary/generative-engine/) composes from several sources.[^geo-paper] **There you are not competing to be the answer, but to be among what the answer uses.** The data only you have (price, stock, shipping time) is in no knowledge base: if your page does not say it, nobody does.


That "answer engine" and AEO are the same thing, or that answer engine means "AI chat". The engine is the system; AEO is the work of optimizing for it, and Google defines it that way, as optimization for answer engines.[^google-ai-guide] And not every answer engine is generative: the Knowledge Graph and Wolfram|Alpha answer from their own data, without composing text from third-party pages.[^google-knowledge-graph][^wolfram-about]


## How to check it

1. **List the ten questions you get asked most** and classify them: fact, definition or recommendation. Only the last two can bring a citation.
2. **Search them on Google** and note what comes up on top: a facts panel, a featured snippet, an AI Overview or just links.
3. **Ask them in AI Mode, in Perplexity and in ChatGPT with search**, on different days, and note the sources each one cites.
4. **Confirm that your pages can be a source on Google**: indexed and with snippets allowed, which is the condition for appearing as a link in its AI features.[^google-ai-features]
5. **Go over which data only you have** (price, stock, shipping, warranty) and check it is written on the page, as text.[^google-ai-features]

## Related concept: generative engine

A [generative engine](/glossary/generative-engine/) is the third family in the table: a system that combines generative models with a search engine and writes its answer from the sources it retrieves.[^geo-paper] **Every generative engine is an answer engine, but not every answer engine is generative**: a facts panel or a knowledge calculator also answers, without composing anything from other people's pages. The distinction matters because only in generative engines and extracts can a store be the source, which is what GEO pursues.

To keep going, the [AI search reading path](/learn/ai-search/) orders the rest of the terms, and [AI search](/glossary/ai-search/) explains how each service retrieves pages.