Search marketing has acquired a new collection of acronyms. Alongside SEO, businesses are now being offered AEO, meaning answer engine optimisation, and GEO, or generative engine optimisation. The suggestion is often that appearing in AI-generated search results requires a new set of rules.
Google's own guidance is rather less dramatic. Its current position is that established SEO practices remain relevant to AI Overviews and AI Mode, and that there are no additional technical requirements or special optimisations needed simply to become eligible for them. Google's guidance for AI features and websites says pages still need to be crawlable, indexed and eligible to appear in normal Google Search.
That does not mean AI search changes nothing. It changes how information can be discovered and presented, and companies should pay attention to it. But there is an important difference between adapting to a changing search experience and rebuilding an SEO strategy around techniques that Google itself says are unnecessary.
AI search is still built on search
AI Overviews are generated summaries that can appear above or among conventional Google results. AI Mode is a more conversational search experience designed for questions involving exploration, reasoning or comparison. Both can direct people towards supporting websites rather than operating as completely separate indexes of the web.
Google explains that these features draw on its existing Search index and ranking systems. They can also perform what Google calls 'query fan-out', where a broader question leads to several related searches behind the scenes. A person asking one complicated question may therefore be connected with pages that answer different parts of it, even when those pages do not repeat the exact wording of the original query.
This helps explain why Google's advice is still recognisable to anyone familiar with good SEO. Make important content accessible to search engines. Use sensible internal links. Keep business and product information accurate. Offer a good experience when somebody reaches the site. Publish material that genuinely answers the needs of the intended audience.
There is no guarantee that following those practices will make a page appear in an AI-generated response, just as conventional SEO cannot guarantee a particular ranking. They establish eligibility and improve the quality of the website rather than providing a secret route into an AI answer.
Several supposed AI-search requirements can be ignored
Perhaps the most useful part of Google's generative AI optimisation guide is its myth-busting section. It deals directly with several techniques that have accumulated around AEO and GEO.
- You do not need a special AI file. Google says it does not use llms.txt or similar AI-specific files to determine visibility in Search. Other services may choose to use such files, but adding one does not improve Google rankings or AI visibility.
- You do not need special AI schema. Structured data remains useful where it supports existing Google search features, but Google says there is no special Schema.org markup required for AI Overviews or AI Mode.
- You do not need to rewrite every page for AI. Google's systems are intended to understand meaning, synonyms and context. Trying to manufacture every possible variation of a question is unnecessary and can produce repetitive, poor-quality content.
- You do not need to split every answer into tiny chunks. Clear structure helps readers, but Google explicitly says there is no requirement to chop content into small sections purely so an AI system can understand it.
This is useful commercially as well as technically. When a new optimisation service or recommendation is proposed, the question should not be whether it sounds plausible. Ask what evidence supports it, which search platform it applies to, what problem it solves and whether the platform itself recommends it.
Content may need to become better, not stranger
Google does place particular emphasis on what it describes as valuable, non-commodity content. In plain English, that means publishing something worth finding rather than another version of information that already exists everywhere else.
A company's own research, detailed product knowledge, original photographs, informed comparisons, useful explanations and genuine subject expertise can all provide substance that a generic summary cannot. Google's guidance specifically warns against simply recycling what other websites have already said or producing material that could easily have been generated from common information.
This does not require every page to contain a groundbreaking piece of research. A service page still needs to explain a service clearly. A product page still needs accurate product information. The opportunity is to identify subjects where the company genuinely knows something useful and make that knowledge accessible rather than producing content simply because a keyword tool says a phrase receives searches.
It is also worth resisting the temptation to create dozens of near-identical pages for possible AI questions. Google warns that producing content at scale primarily to manipulate search rankings or generative AI responses can fall within its spam policies. A smaller collection of useful pages may be a much better investment than a content factory aimed at hypothetical AI prompts.
Measure what is happening before inventing a problem
Another practical change is measurement. Google introduced dedicated generative AI performance reporting in Search Console during 2026 and, as of 31 August, says those insights have been rolled out to websites worldwide. The reports can show impressions, pages, countries, devices and changes over time for visibility within Google's generative AI features.
That gives website owners a better starting point than assuming AI search has either destroyed their visibility or created a huge new source of traffic. Look at the evidence available for the individual website. Search Console data, conventional organic performance, conversions and wider business outcomes can then be considered together.
Google also advises caution around third-party tools claiming access to internal Google ranking or AI metrics. External tools can still be useful for research and monitoring, but their scores should not automatically become business targets simply because they have been labelled 'AI visibility'.
So, does your business need AI SEO?
If 'AI SEO' means making a website technically accessible, publishing genuinely useful material, keeping information accurate, understanding users and measuring search performance, then much of it is already SEO.
If it means buying a collection of AI-specific fixes because conventional SEO supposedly no longer applies, Google's current guidance provides good reason to be sceptical.
AI search is worth monitoring, and the way people discover websites is clearly changing. Roled has previously looked at that broader shift in online discovery. The more useful response now is not to chase every new acronym. Keep the fundamentals working, use the new measurement available, and judge genuinely new techniques by evidence rather than by how convincingly they have been renamed.
