Customer Impact

SEO & GEO

What is a generative search engine?

Copy for AI

A generative search engine answers your question with an AI-written text, not links.

You ask a full question and get a summarised answer back, often with a few sources alongside. That’s what generative search means.

The verdict: in a generative search engine, being named in the answer matters more than ranking for the click. Clear, well-structured content is how you get there.

Below: how it differs from classic search, how it works, which ones you already use, and what it means for your visibility.

What exactly is a generative search engine?

It’s a search system that combines finding information with a language model that writes the answer.

The word “generative” refers to generative AI: technology that creates new text. It draws on what it learned in training and on what it finds online at that moment.

New to the model behind it? Our explainer on what an LLM is covers the basics.

A classic search engine shows you ten blue links. A generative search engine writes an answer in full sentences, often with a few sources.

The difference lies in what you get back:

  • Classic search gives a referral: here are the pages, choose for yourself.
  • Generative search gives a conclusion: this is the answer, and these are its sources.

To the user, it feels like asking a well-informed colleague rather than browsing a library catalogue.

What a generative answer looks like

Here’s an example: we asked ChatGPT to explain generative engine optimization in three bullets.

ChatGPT free answer to our GEO prompt, an example of a generative search engine answer: three bullets with bold lead lines on visibility in AI answers, authority beyond SEO, and AI visibility as a growth channel, with no sources ChatGPT free (temporary chat), 17 October 2026: three bullets, about 90 words, no sources.

What it shows: a finished answer, written for you, with no list of pages to pick from.

How does a generative search engine differ from Google?

A generative search engine gives an answer, while classic search gives a choice.

With the blue links, you decide which page is relevant and click through to it. With a generative search engine, the model largely makes that choice for you.

That has three concrete consequences.

1. You phrase things differently

In Google, you typed loose keywords such as “compare accounting software”.

To a generative search engine, you ask a full question. For example: “which accounting software suits a small B2B company in Belgium”.

Our analysis of ChatGPT queries getting longer shows where that shift leads.

2. You click less

Because the answer is already there, many users no longer move on to a website.

Pew Research found Google users clicked a result in 8% of visits with an AI summary, against 15% without one.

The sample is big. Pew tracked 900 US adults, and 12,593 of their 68,879 Google searches showed an AI summary.

Here’s the click gap side by side:

PEW RESEARCH, 2025 Clicks on a result, share of visits No AI summary 15% With AI summary 8% Browsing data from 900 US adults, March 2025, across 68,879 Google searches. The mention in the answer becomes more important than the click.

The mention in the answer now weighs more than the click. Our piece on AI Overviews and fewer clicks digs into what that does to traffic.

3. Sources are selected, not listed

The model picks a handful of sources. It may cite them.

You don’t see a ranking. You see a verdict.

Watch out: a page that ranks third in Google can still be left out of the answer.

The two worlds overlap

These two worlds exist side by side. Google itself became a generative search engine when it put AI Overviews above the regular results.

The blue links remain below them. So the line between “search engine” and “generative search engine” is blurring.

Want the full difference between the two disciplines? Read our explanation of GEO vs SEO.

Classic searchGenerative search
What you getA ranked list of linksA written answer with a few sources
How you askLoose keywordsA full question
Who choosesYou click and compareThe model selects and summarises
Win for a businessThe clickThe mention in the answer
ExamplesGoogle’s blue linksChatGPT, Gemini, Perplexity, AI Overviews

How does a generative search engine work?

Usually in three steps: it understands your question, pulls in current information, and writes an answer with it.

The middle step is crucial and is called retrieval. The model first looks up relevant sources on the web or in an index, and uses them to support its answer.

Step by step

In practice, it looks like this:

  1. You ask a question.
  2. The system turns it into searches and fetches a set of pages.
  3. The language model reads those pages and filters what’s relevant.
  4. It writes an answer in natural language.
  5. It adds source references, so you can check where the information comes from.

Google calls step 2 query fan-out. AI Overviews and AI Mode may issue “multiple related searches across subtopics and data sources”.

UNDER THE HOOD How a generative search engine builds its answer Step 1 Understands your question turns it into search queries Step 2 Fetches sources retrieval of pages Step 3 Writes the answer with source citation If you don't make it into the retrieval step, you're not in the answer.
A generative search engine understands your question, pulls in current sources and writes an answer with them.

The technical side of that pipeline is in our guide to AI search architecture and RAG.

Why it fetches sources at all

Pulling in current sources stops the model from relying on memory alone. That memory is frozen at the moment of training and can be outdated.

The process by which an AI anchors its answer in verifiable, current sources is called grounding.

How that decides whether your information gets picked up, you can read in grounding: how AI decides what is true.

What it means for your page

This is the crux. If your page isn’t retrieved and understood in that retrieval step, you don’t end up in the answer.

Rule of thumb: no retrieval, no mention.

Which generative search engines are there?

The best-known are ChatGPT, Google Gemini, Perplexity and the AI Overviews in Google Search.

They do the same job. But they make different choices in how they show and weigh sources.

ChatGPT

ChatGPT from OpenAI can search the web directly and support answers with sources.

It has grown into one of the most-used entry points for questions people used to ask Google. Which pages it leans on is in our study of which sources ChatGPT cites.

Google Gemini, AI Overviews and AI Mode

Google Gemini and the AI Overviews bring generative answers to the world’s largest search platform.

Google also has a more extensive AI Mode, in which search becomes fully conversational.

Here’s Gemini answering the same GEO prompt we gave ChatGPT:

Gemini answer to our GEO prompt, a generative search engine reply in dark mode: a one-sentence definition followed by three bold bullets, with no sources Gemini (Google Workspace account, Flash model), 17 October 2026: an intro sentence plus three bullets, about 120 words.

What it shows: the same question gets a differently shaped answer in each engine.

Perplexity

Perplexity positions itself as the answer engine that leans heavily on visible sources.

It searches the web in real time and every response includes citations to the original pages.

You can open the full list of pages behind one answer:

Perplexity sources panel for our GEO prompt listing the 10 web pages behind the answer of this generative search engine Perplexity’s sources panel for our own test prompt, 17 October 2026: 10 web pages behind one answer.

What it proves: a generative answer rests on a small set of pages, and only those pages get the credit.

How to be one of them is in our guide to Perplexity SEO.

Microsoft Copilot

Microsoft Copilot brings generative answers to Bing and the Microsoft environment.

Each engine picks differently

Each of these systems picks its sources in its own way. Being visible in one doesn’t automatically mean being visible in another.

We wrote separate guides on getting found in ChatGPT and on getting found in Google AI Overviews.

For a side-by-side view, see our comparison of AI search engines.

Watch out: this landscape moves fast. Names, features and market shares change often, so check the current situation before you base decisions on it.

Why is a generative search engine important for your business?

Because a growing share of your future customers start their search there. If it doesn’t name your brand, you don’t exist in that answer.

The buyer then builds a shortlist. Without you.

It hits the orientation and comparison phase

Think of someone who asks “which approach suits a B2B company”. Or “what’s the difference between option X and Y”.

They get a summarised answer and often no longer look at ten links.

Named, or invisible

If you’re named in that answer, you’re early in the buyer’s thinking, even before they compare names.

If you aren’t named, the competitor who does appear gets a free assist.

Don’t panic, but don’t wait

One nuance: not every search shifts to AI. For local or strongly purchase-driven searches, classic search remains dominant for now.

So the sensible response isn’t panic. It’s being present in both places now, while many competitors are still waiting.

For B2B companies, we worked this out further in our approach to getting found in ChatGPT.

Quick check: see whether AI engines already mention you with our free AI visibility checker.

How do you become visible in a generative search engine?

Through content that answers a real question directly and clearly, structured so a model can read and cite it.

This field is called generative engine optimization. The full approach is in our complete guide to GEO.

The basic principles are recognisable.

Answer the question at the top

Put the answer in the first sentence of a section, not after three introductory paragraphs.

Models pick up that direct phrasing faster. Our guide on how to write content AI cites shows the pattern.

Structure for readability

Use headings that are real questions, short paragraphs and lists. What a human scans easily, a model extracts easily.

For Google, no special markup or AI text file is needed.

It says there are no additional requirements to appear in AI Overviews or AI Mode beyond being indexed and eligible for a snippet.

Build authority

A brand that’s mentioned in several trustworthy places is chosen as a source more often.

Mentions and consistency beat tricks.

Be concrete and factual

Definitions, clear figures in context and unambiguous wording give a model something to hold on to when it cites you.

The common thread

Don’t write for the machine, but for the reader. Then make it as easy as possible for the machine to pick up your clear answer.

It’s not a bag of tricks. It’s good, reliable content, built a bit more sharply.

Frequently asked questions

Is Google a generative search engine?

Partly. Google Search now shows AI Overviews and offers AI Mode, which write answers, while the classic blue links stay below them.

What are examples of generative search engines?

ChatGPT with search, Google Gemini, Google AI Overviews and AI Mode, Perplexity and Microsoft Copilot.

Do generative search engines cite their sources?

Most do, in different ways. Perplexity includes citations in every response; ChatGPT, Google and Copilot often show a few source links next to the answer.

Do I need special markup to appear in generative search?

Not for Google: it says there are no additional requirements beyond being indexed and eligible for a snippet. Clear, well-structured content matters more.

The short summary

A generative search engine gives a written AI answer instead of a list of links. It bases that answer on sources it retrieves and weighs itself.

ChatGPT, Gemini, Perplexity and Google AI Overviews are the best-known examples.

For your business, the goal shifts from winning a click to being named in the answer.

You win that with clear, well-structured content and a brand that shows up in several places.

And that visibility only counts once it turns into leads and revenue.

Want to know whether your business already shows up in these answers, and where your opportunities lie? Contact us and you’ll hear within 24 hours where you stand.

Free website scan

Enter your website and get an automatic scan within minutes, with concrete technical and SEO improvements. No sales pitch.

Where should we send your report?

We only use your details for your scan. No spam, unsubscribe anytime.