Two definitions, in plain words
Search engine optimisation is the work of getting a page ranked in a list of results so that a person clicks it. The unit is the page. The measure is the click.
Answer engine optimisation is the work of getting named as a source when the engine writes the answer itself. Google's AI Overviews and AI Mode, ChatGPT, Perplexity, Copilot and Claude all do this. The person reads a paragraph, sees a short list of sources, and usually does not click any of them. The unit is the paragraph. The measure is whether you are in the list.
Some vendors call the second thing GEO, generative engine optimisation, or AI SEO. Some sell all three as separate products. It is one job. I sell it once, under one name, as SEO and AEO.
Why the distinction matters now
The click is worth less than it was. Pew Research Center, working from March 2025 browsing data across about 900 US adults and roughly 70,000 searches, found that when an AI Overview appeared, click-through on the results below it fell from about 15% to about 8%. Links inside the AI answer were clicked about 1% of the time.
So a page can hold its ranking and lose half its clicks. A page can be cited in the answer and get almost no clicks at all. Ranking still matters, because the engines retrieve from the same index. It is no longer the only measure.
What is the same work
Most of it.
- Crawlability. If a crawler cannot fetch the page, neither kind of engine can use it. Check robots.txt for AI crawlers blocked by accident.
- Speed and indexing. Same as ever.
- Titles, headings and internal links. The engines read the same HTML.
- Consistent facts. Name, address, locations, services, the same on every page and in every listing. An engine that finds three different location counts trusts none of them.
- Pages that answer the question a buyer typed. That was always the job.
What is different
Two things. Both change how a page is written rather than which pages exist.
Retrieval scores paragraphs. An answer engine pulls candidate passages, ranks them, and writes from the top few. A page that buries its answer in paragraph nine can rank well as a page and never be lifted as a passage. The fix is a question-shaped heading followed by a paragraph that answers it in full, with the number in it, before any preamble.
Citations often go to third-party pages. When an engine answers a question about your category, the sources it names are often not the businesses in that category. They are directories, comparison pages, forums and publications that already describe the category. Being cited can mean being described accurately on the pages the engine already trusts. That is off-site work.
Schema is not one of the differences. Google's own AI features documentation says there is no special schema.org structured data you need to add for AI Overviews or AI Mode.
What it costs here
I do not sell them separately, because the work is not separate.
- AI search tracking, A$249 a month. Your client questions checked weekly in five AI tools, with a monthly summary.
- Answer Engine Audit, A$1,500. A one-off check of how five AI tools see your business, with a fix list and a re-test at 60 days.
- Packages at A$2,900, A$4,900 and A$9,500 a month. SEO and AEO in all three, in quarterly terms. Package 3 includes backlinks.
All prices are plus GST. The full list is on the pricing page. The way the number is measured is on how I measure.