Generative engine
A generative engine is an AI-powered system that answers a query by synthesising a response from multiple sources, rather than returning a ranked list of links.
ChatGPT, Perplexity, Google AI Overviews, and Copilot are generative engines. Because they present one synthesised answer — often citing only a handful of sources — visibility in them is closer to winner-take-most than a traditional results page, which raises the value of being the clearest source on a topic.
Generative engines are powered by large language models. To understand how they read, retrieve, and decide what to cite, see our full guide: What Are Large Language Models (LLMs)?
How it works
A generative engine usually rewrites the user's question into one or more searches, retrieves candidate pages or passages from a web index or its own crawl, and passes the best of them to a language model with instructions to answer and cite. Some engines answer from the model's training alone for simple questions and search only when they need current or specific information.
The sources an engine retrieves, and how easily it can extract a clear claim from them, decide who gets named in the answer.
Example
Ask Perplexity which studios in Prishtina build booking systems, and it runs searches, reads a handful of pages and directory listings, and writes a short answer naming a few businesses with links. A studio whose own site never mentions Prishtina or booking systems is unlikely to be one of them.
Common mistakes
- Assuming ranking first on Google guarantees a mention in generated answers.
- Publishing key facts only in images or scripts that engines cannot read.
- Never checking what the engines actually say about your business.
