What is an LLM (and an answer engine)?
TL;DR
An LLM (large language model) is the AI behind tools like ChatGPT, Claude, and Gemini — trained on huge amounts of text to understand and generate language. An answer engine is a product (ChatGPT, Perplexity, Google's AI Overviews) that uses an LLM to answer a question directly, often citing the sources it drew from.
What it is
An LLM is like an extraordinarily well-read assistant: it has read an enormous slice of the internet and predicts, word by word, the most helpful response to what you ask. It doesn't "look things up" the way a database does — it draws on patterns learned in training, and increasingly on live content it retrieves at question time.
An answer engine wraps that ability in a search-like experience: instead of ten blue links, you get a written answer — and, more and more, a short list of the sources behind it. That "cited sources" part is where your website comes in.
Why it matters
- They read the web through machines. LLMs and their crawlers parse pages far better when the facts are labeled — via structured data and files like llms.txt — than when they're buried in messy HTML.
- Citations are the new clicks. Being the source an answer engine quotes is the AI-era equivalent of ranking on page one.
- Access is a choice. robots.txt lets you decide which AI crawlers may read you at all.
How MetaPilot handles it
MetaPilot makes your pages easy for LLMs to parse — accurate structured data, a fresh llms.txt, and clean edge delivery — and shows you which AI crawlers are actually visiting, so you can see and shape how answer engines read your site. What is MetaPilot? →
Related
FAQ
Do LLMs actually read my website?
Their crawlers can (GPTBot, ClaudeBot, PerplexityBot, and others), and answer engines retrieve live pages to cite. You can allow or block them in robots.txt.
How do I get cited by an answer engine?
Make your content easy to parse and trust: clear structure, accurate structured data, freshness, and access for the crawlers you want. That's the heart of AEO.
Is training the same as being cited?
No — training uses content to build the model; citation happens when an engine retrieves and references your page in an answer. You can feel differently about each, and control crawlers accordingly.
Ask an AI to explain this