Search engine optimisation (SEO)
In short
Search engine optimisation is not a separate service from answer engine optimisation — it is the layer underneath it. Answer engines retrieve heavily from indexed web content, so a page classic crawlers cannot reach, read or rank is a page they cannot cite either. What changed is that ranking stopped being sufficient: Ahrefs found in March 2026 that 38% of pages cited in Google's AI Overviews also ranked in the top ten for the same query, down from 76% in July 2025. This page covers crawlability and rendering, indexing reality, heading structure, Core Web Vitals against Google's published thresholds, and exactly where SEO stops and something else has to take over.
Is SEO a different service from AEO and GEO?
Not a different service — a different layer of the same one. Answer engine optimisation (AEO), generative engine optimisation (GEO) and AI search optimisation are three names for one discipline. Search engine optimisation — search engine optimization, if you spell it the American way — is what that discipline stands on. Answer engines retrieve heavily from indexed web content, so a page classic crawlers cannot reach, read or rank is a page those engines cannot cite either.
This page is the honest version of the sales argument: it explains what SEO still decides, and it says plainly where SEO stops and something else has to take over. If you want the combined service, it is here.
Has AI search made SEO obsolete?
No — but it has stopped being sufficient, which is a different claim and an easier one to test. Ahrefs found in March 2026 that 38% of pages cited in Google's AI Overviews also ranked in the top ten for the same query, down from 76% in July 2025. Read that in the direction people usually miss: the overlap fell, it did not vanish. Roughly two in five cited pages were still top-ten performers.
So ranking still correlates with being cited, and it no longer determines it. A company that abandons search engine optimisation to chase AI visibility removes the floor it was standing on. A company that does only search engine optimisation and assumes citations follow is relying on a relationship that has weakened measurably in under a year.
What does the SEO layer actually decide?
Whether the content exists, as far as a machine is concerned. Four things in particular:
| Check | What fails when it is wrong |
|---|---|
| Crawlability and rendering | The crawler receives an empty or partial document, so there is no passage to retrieve at any price |
| Indexing reality | Pages you believe are live are excluded, canonicalised away or duplicated against each other, so the wrong URL competes |
| Information architecture and headings | Content exists but is not segmented into passages a retrieval step can isolate and reuse |
| Structured data | Markup contradicts the visible page, which is worse than having none — it is an accuracy signal pointing the wrong way |
Do AI crawlers read pages the way Googlebot does?
Mostly not, and this is the sharpest practical difference. Vercel and MERJ tracked over 500 million GPTBot requests in December 2024 and found no evidence of JavaScript ever being executed — GPTBot requested JavaScript files 11.5% of the time and ClaudeBot 23.8%, and neither ran them. Gemini and Applebot are the exceptions, because they reuse rendering infrastructure that already exists.
Googlebot renders JavaScript. Most AI crawlers do not. That means a client-rendered site can rank perfectly respectably in classic search while showing an effectively blank page to the systems now answering your buyers' questions. Frameworks such as Next.js, Nuxt, SvelteKit and Astro all have server-rendered or static modes that resolve this without a rewrite, so the first question is which mode you are actually in, not whether you need to rebuild.
What are the current Core Web Vitals thresholds?
Google's published thresholds are Largest Contentful Paint at or under 2.5 seconds, Interaction to Next Paint at or under 200 milliseconds, and Cumulative Layout Shift at or under 0.1 — each measured at the 75th percentile of field data. First Input Delay was retired on 12 March 2024 and replaced by Interaction to Next Paint, so any tool still reporting FID is out of date and any agency still quoting it has not looked recently.
Where does SEO stop and answer engine optimisation begin?
At the point where the question changes from "can this page be found and ranked?" to "can this passage be lifted out, understood alone, and attributed to the right company?" Those need different work:
- Entity resolution — search ranks documents; answer engines attribute claims to organisations. If a system cannot establish which company you are, improving your pages strengthens somebody else's entity.
- Passage architecture — a page can rank well while every section depends on the one above it. Retrieval takes the section, not the page.
- Corroboration — rankings are earned on your own site. Citations lean substantially on third-party sources describing you, which are not yours to edit directly.
- Measurement — position tracking answers a different question from answer share and citation rate, and neither substitutes for the other.
The full sequence, and why the order is load-bearing, is on the method page.
Do I need SEO and AEO as two engagements?
No. It is one piece of work with a sequence: the search layer gets checked and fixed first, because everything above it depends on a crawler reaching and reading the page. There is no point optimising passages on a page that is not being retrieved at all. Being quoted separately for "SEO" and "AEO" and "GEO" is a pricing structure, not a technical one.
What to do about your own site
Test the mechanical layer before commissioning any content work: request one of your pages with a crawler user-agent and no JavaScript execution, then search the returned HTML for a phrase you know is on the page. If it is missing, nothing above this layer matters yet.
That test is included, along with the per-engine visibility results, in the free check.
Frequently asked questions
Has AI search made SEO obsolete?
No. Answer engines retrieve from indexed web content, so a site that classic crawlers cannot read or rank is a site those engines cannot cite either. What has changed is that ranking alone no longer buys a citation: Ahrefs found 38% of pages cited in Google AI Overviews also ranked in the top ten for the same query in March 2026, down from 76% in July 2025.
What are the current Core Web Vitals thresholds?
Largest Contentful Paint at or under 2.5 seconds, Interaction to Next Paint at or under 200 milliseconds, and Cumulative Layout Shift at or under 0.1, each measured at the 75th percentile of field data. First Input Delay was retired on 12 March 2024 and replaced by Interaction to Next Paint, so any tool still reporting FID is out of date.
Do I need SEO and AEO as two separate engagements?
No. They are one piece of work with a sequence. The SEO layer gets checked and fixed first because everything above it depends on a crawler being able to reach and read the page, and there is no point optimising passages on a page that is not being retrieved at all.
How is progress reported?
Technical fixes are reported as a before-and-after against the specific check that failed, with the measurement method stated. Visibility is reported as answer share and citation rate on a fixed prompt set re-run across engines. No projection of future rankings or traffic is included, because those would be invented numbers.
Find out what the engines currently say about you
Send me your company name and website. I run a set of buyer questions across ChatGPT, Perplexity, Gemini, Claude and Google's AI Overviews, and send back a recorded walkthrough of what came out: where you appeared, where you did not, who was named instead, and the two or three structural reasons why. No charge, no obligation, and you keep the findings whether or not you decide to hire me.
I run these myself, so there is a queue. Expect a few working days rather than an instant report.