[ai overviews & aeo]
More of your buyers now ask an AI tool for a recommendation instead of scrolling a results page. If your name is not in that answer, you are not on the shortlist. Ranking third does not help you there.
Classic SEO aims for a click. AI search aims for a citation. Those are different targets and they reward different things.
Ask a model which prop firms allow news trading and it does not hand over ten links. It writes an answer, names a few firms, and cites a handful of sources. You are either in that answer or you are not. There is no page two.
The names that come up are often not the sites ranking first. Review sites, forums and mid-sized publishers get cited ahead of the firm’s own site more often than you would expect.
The model needs to know what your brand is, what category it sits in and who runs it. If that is fuzzy it picks someone it is sure about. Background in entity SEO.
Direct answers, stated early, with the actual number or rule included. Content that buries the answer under three paragraphs does not get picked up.
Models favour sources that commit to detail. Drawdown percentages, splits and fees as real numbers. Marketing adjectives are no use to them.
Schema that describes what is actually on the page, so the content is easy to read machine-side.
Consistency across review sites, forums and press. AI answers lean on agreement, so one self-published claim carries little weight.
If your rules changed and the web still shows the old ones, the old ones get quoted. This is very fixable and often ignored.
I run a set of buyer questions across the main assistants and record what comes back. Whether you are named, who else is, what is cited, and where the description is simply wrong. Most firms have never looked, and the results are usually a surprise.
Not every question is worth chasing. The ones that count are the ones asked right before someone chooses.
Sometimes the fix is on your site, and means restructuring pages so the answer can be lifted. Often it is off your site, where a widely cited page has old information about you. Correcting that does more than any on-page change.
Specific, factual pages on the questions your buyers ask. This overlaps with normal SEO, but the formatting priorities are different.
The same questions, run again, so you can see movement. It is the only honest way to report on this, because there is no rank tracker for AI answers yet.
AEO, GEO, AI search optimisation. The labels are new and the ground is still moving. Nobody has five years of data, because these surfaces did not exist five years ago.
What I can say is that clear definition, specific facts and consistent description are working now. Measuring where you stand costs very little next to finding out in a year that your rivals own the answers. I wrote up what I have seen in how prop firms get cited in AI Overviews.
Send me your brand and your main rivals. I will run the questions and show you what comes back.
There is a lot of overlap, and good SEO helps. But the target is different. Ranking wins a click. Being cited wins a mention. A page built to win a click is often not built to be quoted.
No, and I would be careful with anyone who says they can. These systems change without notice. What you can control is being clearly defined, factually specific and described the same way across the web.
A fixed set of buyer questions, run across the main assistants, recorded before and after. I track whether you are named, who else is named, and what gets cited. It is done by hand because the automated tools are not reliable yet.
Often better. Operators researching a payment provider or a risk tool ask assistants detailed questions, and very few vendors show up in those answers yet.
That is common, and it is the most useful thing to fix first. It usually traces back to one outdated page that keeps getting reused. Find it, correct it, and the answers tend to follow.