For twenty years the game was simple: rank on page one, win the click.
That game is changing. More and more people are asking ChatGPT, Gemini, Perplexity or Claude instead of typing into Google. They don’t get ten blue links. They get one answer. If your business isn’t in that answer, you don’t exist.
That’s Answer Engine Optimisation (AEO). And if you care about SEO, you need to care about it now.
SEO vs AEO in one line
- SEO: get found in a list of results.
- AEO: get cited in the only result.
Same goal (be the source people find), different mechanics. Search engines reward keywords, backlinks and authority signals. Answer engines reward content that’s clear, structured, factual and easy for a model to lift and quote.
The upside is bigger. A top-three Google ranking still splits traffic with nine other sites. Being the source an AI cites is closer to a monopoly on that question.
“Can’t I just buy my way in?”
You can try. ChatGPT ads launched in early 2026 and self-serve access opened in May. Google is doing the same in AI Overviews. The money is flowing in.
The results aren’t flowing back. Here’s the problem in three parts.
1. Nobody knows what “good” looks like. OpenAI hasn’t published cross-advertiser benchmarks for click-through rate, conversion rate, cost per acquisition or return on ad spend. Every advertiser is guessing at what their numbers should be.
2. You can’t measure it properly. The platform is privacy-first by design, which makes user-level attribution weak from day one. Third-party measurement has been promised but no partners or timeline have been named. Dashboard conversions tell you what the platform tracked, not what you’d have got without the spend. The only way to prove it works is running holdout experiments, and most businesses aren’t set up for that.
3. The inventory is unstable. Ad frequency has swung from a quarter of replies, to almost zero, to over half within weeks. Pricing has compressed sharply since launch. Coverage is heavily US-weighted. Some categories get ads on nearly every relevant prompt; others get none. You’re buying into a market that hasn’t found its floor.
Paid AI placement will mature. Right now it’s a bet with no scoreboard.
Where the money actually is
Organic AI visibility. The answer the model gives before it appends a sponsored link.
That answer is built from content it trusts. Getting into it costs time, not media budget, and the return compounds. A page that’s cited today is still cited next month. An ad stops the moment you stop paying.
What AEO actually looks like
Most of it is good SEO done properly, plus a few things search never forced you to do:
- Answer the question in the first sentence. Models lift direct answers. Don’t bury them under 400 words of preamble.
- Structure for extraction. Clear headings, short paragraphs, FAQs, tables. Make your content easy to quote.
- Be specific and factual. Numbers, names, dates, definitions. Vague content doesn’t get cited.
- Use schema markup. FAQ, HowTo, Product, Organisation. It’s how machines confirm what your page is about.
- Build entity authority. Consistent business details everywhere. Mentions on sites the models already trust. Reviews, directories, industry publications.
- Cover the whole question, not just the keyword. Answer engines pull the most complete source, not the one with the best keyword density.
- Keep it current. Models favour recently updated content for anything time-sensitive.
The honest bottom line
SEO isn’t dead. Google still sends more traffic than any AI tool. But the share of questions being answered without a click is growing every quarter, and paid placement in that space can’t yet prove its worth.
The businesses that win the next five years will be the ones the answer engines already trust. That’s earned, not bought.
Start earning it now.
