For more than two decades, the basic model of search has been remarkably consistent: somebody types a query into a search engine, receives a list of results and clicks through to a website.
Businesses consequently invested heavily in SEO to appear as high as possible on that list.
AI is beginning to change that relationship.
Search engines can now answer increasingly complex questions directly. Google AI Overviews and AI Mode synthesise information from multiple sources, while standalone AI assistants have accustomed users to asking questions conversationally and receiving an immediate response rather than selecting between ten blue links.
For businesses, this creates an important distinction.
Being found no longer necessarily means being clicked.
And that means the traditional definition of SEO success needs to evolve.
From search results to answers
Traditional search was primarily a retrieval system. A search engine identified pages that appeared relevant and presented them to the user.
AI-powered search increasingly adds another layer: synthesis.
Rather than simply identifying sources, the system can extract information from multiple places and construct an answer itself.
Google describes its AI search features as capable of using a “query fan-out” approach, conducting multiple related searches across subtopics and sources before constructing a response. That means a business can potentially become a source for an AI-generated answer without occupying the traditional number-one organic position for the original query.
That is a significant change in how digital visibility works.
The click is no longer guaranteed
The commercial tension is obvious.
If an AI system can satisfactorily answer a question on the search results page, why does the user need to visit the source?
Early research suggests this is already affecting behaviour. A 2026 study examining Google searches from a panel of 900 US adults found that users clicked sources cited within AI Overviews in only around 1% of visits to those Overviews. The researchers also found that AI Overviews were associated with fewer clicks and a greater likelihood that the browsing session ended on the search page.
For publishers, comparison businesses and organisations that historically relied heavily on informational search traffic, this presents a structural challenge.
A page could still be useful to a search engine — perhaps even useful enough to inform its AI-generated answer — without generating anything approaching the traffic it once did.
Is traditional SEO dead?
No.
In fact, many of the foundations of good SEO may become more important rather than less.
Google explicitly says that the established fundamentals of SEO continue to apply to AI Overviews and AI Mode. Pages still need to be crawlable, indexable and understandable, with important information available as text and sensible internal linking helping systems discover content.
Technical SEO therefore does not disappear.
Neither do content quality, site architecture, links or authority.
What changes is the objective.
Traditional SEO has typically concentrated on questions such as:
Can the page rank?
For which keyword?
In which position?
How much organic traffic does that position generate?
The emerging model introduces additional questions:
Does an AI system understand what the organisation does?
Does it recognise the organisation as credible on the subject?
Is its information sufficiently clear to extract?
Is the business being cited or surfaced when relevant questions are asked?
And, perhaps most importantly, does that visibility ultimately influence a commercial decision?
SEO, AEO and GEO
This shift has created a growing collection of acronyms.
SEO — Search Engine Optimisation — remains the established discipline of improving visibility within search engines.
AEO — Answer Engine Optimisation — generally describes structuring information so that systems can understand it and use it to answer specific questions.
GEO — Generative Engine Optimisation — takes that idea further, focusing on visibility within responses produced by generative AI systems.
The terminology is less important than the underlying change.
Businesses are moving from optimising purely for rankings towards optimising for machine understanding, retrieval and recommendation.
That does not mean creating an entirely separate “AI website”. In Google’s case, there is no special AI schema or machine-readable file required to appear within AI Overviews or AI Mode.
Instead, businesses need to make their existing digital presence substantially clearer.
Structured content becomes more valuable
A surprisingly large number of enterprise websites remain difficult to understand.
Products are described differently across departments. Service pages use internal terminology that customers would never use. Important information sits inside PDFs. Corporate websites contain overlapping pages written at different times by different agencies.
Humans can sometimes work around that inconsistency.
Machines may struggle to establish which information is authoritative.
Businesses should therefore think about structured content at two levels.
The first is technical structured data. Schema and other machine-readable information can help systems understand entities, products, organisations and relationships. Google recommends ensuring structured data accurately reflects the content visible on the page.
The second is the structure of the content itself.
Clear headings, concise definitions, factual product information, comparison tables, FAQs where appropriate, consistent terminology and well-organised information architecture all make content easier for both humans and machines to interpret.
The objective is not to write robotic content for an LLM.
It is to remove ambiguity.
Authority becomes the harder problem
There is a temptation to treat GEO as the next technical SEO checklist.
Add some schema. Rewrite headings as questions. Create hundreds of pages answering long-tail queries. Wait for the AI citations.
That is unlikely to be a durable strategy.
AI systems need to decide not only what information exists but which information they should trust sufficiently to surface.
Microsoft has explicitly described trusted and authoritative information as an important component of AI-powered search, while Google continues to emphasise helpful, reliable and original content.
Authority is therefore becoming a wider organisational challenge.
A business may need strong first-party content, independent coverage, recognised experts, credible citations, consistent information about its products and services, customer evidence and an identifiable relationship with the topics for which it wants to be known.
This starts to blur the boundaries between SEO, PR, content, brand and reputation management.
That is probably a healthy development.
Commodity content has a problem
AI also changes the economics of producing content.
Generating another generic article explaining “ten benefits of cloud computing” has never been easier.
That also makes such an article less distinctive.
Google’s guidance warns against generating large numbers of pages with little additional value, noting that scaled AI-generated content can breach its spam policies when it is produced primarily to manipulate rankings.
Businesses therefore need more information that cannot simply be recreated by asking an AI model to write about the same subject.
That might include proprietary data, original research, genuine technical expertise, customer experience, case studies, specialist analysis or insight derived from operating within a particular market.
The competitive advantage increasingly lies in possessing something worth retrieving.
Measurement needs to change too
Organic sessions and keyword positions will remain useful metrics, but they will tell less of the story.
Google has already begun testing dedicated Search Console reporting for visibility within generative AI search features, reflecting the growing need for businesses to understand how their content appears within these environments.
Digital teams increasingly need to consider a broader set of signals: AI citations, brand mentions, branded search demand, referral quality, conversions, share of relevant answers and whether the organisation appears when prospective customers ask high-value commercial questions.
Traffic remains important.
But visibility without a click can still influence a later decision.
The measurement challenge is proving that influence.
The bigger strategic question
For leadership teams, AI search should not be treated simply as the SEO department’s next optimisation project.
It raises a broader question about how the organisation represents itself digitally.
Can machines understand the business?
Can they distinguish its products from competitors?
Can they identify its expertise?
Is information consistent across websites, platforms, product feeds and external sources?
And when an AI system is asked to recommend, compare or explain something in the organisation’s market, is the company part of the answer?
Traditional SEO is not ending.
But the period in which ranking a webpage and generating a click represented the complete definition of search visibility probably is.
The next phase combines technical SEO, structured information, authority, original content, brand signals and an understanding of how AI systems retrieve and synthesise information.
For enterprises, that makes search less of a marketing channel and more of an information architecture problem.
Silicon Dales works with organisations on enterprise SEO, AI, complex digital platforms, data and technical strategy, helping businesses build digital infrastructure that can be understood by both the people using it today and the increasingly intelligent systems navigating it tomorrow.