Answer engine optimization (AEO) is the practice of getting your brand mentioned and recommended inside AI-generated answers. Where SEO aims to rank a page in a list of links, AEO aims to be named in the answer that replaces the list.
You'll also see GEO — generative engine optimization. The two terms are used interchangeably; nothing meaningful separates them.
Why it's a different problem from SEO
Search gave you a ranked list. Ten links, and a click-through rate you could measure per position.
An AI assistant gives one answer. Someone asking "what are the best running shoes for marathon training?" gets a shortlist of three or four brands and a recommendation. You're either in that answer or you aren't — there is no page two, and no consolation traffic from position seven.
Three consequences follow, and they're what make AEO its own discipline:
Mentions matter more than rankings. Being named in the answer is the win. Your page ranking well matters only insofar as it makes that mention likelier.
Pages you don't own decide your fate. Assistants synthesise from multiple sources, so a magazine roundup or a forum thread often carries more weight than your own site. Being absent from the roundup everyone cites is a bigger problem than any on-site issue.
There's no rank tracker. Answers vary between identical questions, so there's no stable position to check. The unit of measurement becomes a rate: how often you're mentioned across many checks.
What actually moves it
Roughly in order of impact:
Being mentioned on sites assistants trust — roundups, reviews, comparisons, forums. The highest-leverage and slowest work.
Having quotable content — specific claims, real numbers, expert statements. Assistants reach for text they can lift.
Answering the question directly — the answer in the first sentences, not after four paragraphs of preamble.
Being current — recency matters in categories where products and prices change.
Consistent naming — being described the same way everywhere makes you easier to recognise as one entity.
Notably absent from the top: technical markup. Structured data helps machines parse your pages and is worth having, but assistants extract from visible text — it's hygiene rather than a lever. Anyone selling schema as an AI visibility strategy is selling you the least effective part.
Also worth knowing what works against you: keyword stuffing measurably reduces the chance of being cited. Writing for a keyword density algorithm produces exactly the kind of text an assistant skips.
Where to start
Measure first. You can't tell whether you have an on-site problem or an off-site one until you know which sources are shaping the answers in your category — and those two problems need completely different work.
See What makes a page get cited by AI for the detail, and Why AI answers change between platforms and days for why a single spot check won't tell you where you stand.
