Digital marketing and AI · 4 min read

SEO, AEO, GEO: Where the Money Goes

If I had one more marketing dollar today, it would go into the website, the content and the evidence behind them, not into tricks aimed at one machine.

By

Every agency now sells three things. SEO is optimisation for search engines. AEO, answer engine optimisation, is about being the answer an assistant gives. GEO, generative engine optimisation, is about being cited by systems that write their own responses. The sales pitch usually implies you have to choose, or buy all three separately.

I would not choose one and ignore the others. Traditional search is still the foundation. On top of it, I want the same underlying content, evidence, structure and authority to be understandable and retrievable by answer engines. And I would put money into that underlying system before I paid anyone for an “AI ranking” hack.

Google, for its part, has said much the same about its own products. Its 2026 guide to generative AI features in Search says plainly that “optimizing for generative AI search is optimizing for the search experience, and thus still SEO.”

Here is how I would split a budget.

Foundational: pay for this first

A site the machines can read. Complete HTML from the server, fast pages, clean addresses, nothing important hidden behind scripts, images or PDFs. Google’s AI features have “no additional requirements” beyond being indexed and eligible for a snippet. Other assistants run their own crawlers, and several of them do not execute JavaScript. Get this wrong and nothing else on this list works. I have written about it separately in What a Crawler Actually Receives.

Content that is not a commodity. Google’s guide asks for “a unique point of view” and warns against recycling what is already everywhere. Every answer engine has the same problem: if your page says what fifty others say, there is no reason to cite yours. What only you can publish is your own data, your methodology, your prices, your cases, your expertise with your name on it.

Evidence. Claims that can be checked: who you are, what you have done, for whom, since when, with proof on the page. This is where most companies are thin. It is also the part that helps a human buyer just as much as a machine.

Incremental: worth paying for once the foundation is there

Structured data that matches the page. Schema markup describing the organisation, people, products, reviews and articles. Google is clear that it “isn’t required for generative AI search,” and I would not buy it as an AI trick. I would use it because it states facts and relationships unambiguously, and it has to match what is visible.

Being present where the answers come from. Your business profiles, the directories and comparison platforms in your category, the publications your buyers read. Earned, accurate mentions, not manufactured ones. Google lists inauthentic mentions among the things that do not work.

The other engines’ plumbing. Make sure your robots file allows the search crawlers you want, such as OAI-SearchBot, Claude-SearchBot and PerplexityBot, and blocks only what you mean to block.

Measurable: what I would actually watch

  • Search Console. Traffic from AI Overviews and AI Mode is counted in the ordinary Performance report, and in June 2026 Google introduced dedicated generative AI performance reports.
  • Referral traffic in analytics from the assistants themselves.
  • Enquiries and sales by source. A citation that never produces a visit or a lead is a vanity number.
  • What the assistants actually say when asked about your category and your company. Check it by hand, regularly. It is free and it is the most direct test there is.

Hype: what I would not pay for

  • Anything guaranteeing a position in an AI answer. Nobody controls that, any more than anyone could guarantee a first-page ranking.
  • llms.txt as a ranking lever. It is a proposal, not a standard. This site publishes one, because it costs nothing. Google says you do not need it, or “AI text files” of any kind, to appear in its AI features. Do not pay a retainer for it.
  • Rewriting everything “for AI.” Google says there is no need to chunk content into tiny pieces or write in a special way for generative search. Clear writing for people is the format.
  • Volume content. Producing pages has never been cheaper, which is exactly why mass-produced pages are worth so little.

The short version

Spend on the thing all three acronyms are reading: a fast, readable website, content only you can write, and evidence a buyer can check. Then make sure every engine can get to it. If an agency’s AEO or GEO proposal does not start there, it is selling you the wrong thing.

Deciding that split for a specific business, rather than in the abstract, is the digital marketing and AI work.

Sources