What answer engine optimisation actually is (and what it isn't)
Answer engine optimisation is the practice of increasing how often a brand is named and cited inside AI-generated answers. It differs from SEO in its unit of success — inclusion in a synthesis rather than a position in a list — and in its unit of measurement, which is a rate across repeated samples rather than a single rank.
The definition, without the marketing
AEO is making a brand more likely to be named and cited inside AI answers. That's it. Everything else — GEO, LLMO, AI SEO — is the same discipline under a different label, and arguing about the acronym is the least productive conversation in this category.
What makes it a distinct discipline is not the tactics, most of which overlap with good SEO. It's two structural differences.
Inclusion replaces ranking. There is no position two inside a paragraph. A page can rank third and never be cited, or rank eleventh and be quoted in full. The question changes from "where are we?" to "are we in, and how often?"
Measurement becomes statistical. Ask the same question twice and you can get two different answers. Any AEO claim without a sample size behind it is a claim about one coin flip.
What AEO is not
It isn't a schema checklist. Structured data is cheap and worth doing, but our own testing on Chinese engines found zero positive correlation between schema density and mention rate, and controlled publishing tests on Western engines put the effect inside the noise band. Anyone selling schema as the AEO lever is selling the easiest thing to invoice for.
It isn't llms.txt. Publish one; it costs an hour. Do not expect it to move anything. Adoption by the major engines is unconfirmed and our controlled tests found no measurable citation effect.
It isn't prompt injection. Hiding instructions in your page for assistants to read is hidden keyword stuffing wearing a new hat, and it ends the same way.
What actually correlates
From our own data and from published research, the signals that show up repeatedly:
- Being cited by sources that are already cited. An inbound link from a domain the engines already trust cut time-to-first-citation by 4.2x in our publishing study.
- Encyclopedic presence. On Chinese engines a Baidu Baike entry was the single strongest predictor of mention rate we measured.
- Third-party description. Brand-owned pages are only about 15% of citations on commercial prompts. The other 85% is reviews, listicles, publishers and community threads describing you.
- Self-contained passages. Retrieval works at chunk level, so a paragraph that only makes sense in sequence is far less quotable than one that stands alone.
The uncomfortable summary
How to know whether it's working
Baseline before you change anything, sample enough to have an interval, resample after, and only call the change real when the intervals stop overlapping. That discipline is unglamorous and it is the entire difference between AEO as a practice and AEO as a pitch.
See this for your brand
Nostimates shows you how your brand shows up across twelve AI engines and search — in a dashboard, or as data in your own tools. Tell us what you want to measure.