Is SEO dead? How AI affects SEO – what the data shows, and what still decides in 2026

SEO has been through a turbulent year. The debate about how AEO (also known as AI SEO or GEO) affects SEO has swung between overly gloomy readings and unrealistically rosy claims – often alongside simplistic or downright wrong advice on how to "optimise for AI citations". So is SEO dead?

At SEOS Design we have chosen to stick to established best practice and base our decisions on reliable data and insights from sources we trust. Now that AEO is no longer brand new, there is enough material to get a clearer picture of what works – and what does not (read: educated guesses – this is SEO, after all).

TL;DR – what is really happening to SEO and AEO?

• SEO is still the foundation. It makes sure your content can be found, and it also forms the database AI systems draw on to generate answers.
• AEO weighs heavily in what gets chosen. AI answers increasingly influence which brands get recommended – especially close to the buying decision.
• SEO and AEO depend on each other. Without SEO, AEO lacks a stable foundation, and without AEO, SEO loses part of its business value. Source

What the data shows:

• The often-cited "SEO collapse" is not supported by broad data. An analysis by Graphite shows that organic search traffic has fallen by only about 2.5% year on year on average. Source
• AI Overviews can reduce CTR on organic results by around 35% when they appear, but they only appear in about 30% of cases. Source
• Traffic from AI citations usually has considerably higher purchase intent. Webflow has reported a 6× higher conversion rate compared with non-brand organic search traffic (Webflow's own figures). Source

What to expect going forward (2026):

• SEO and AEO are moving from individual keywords to whole ecosystems of questions around a topic.
• Long-tail questions are becoming longer, more conversational and more commercially focused.
• Because AI search has no clear search volume data, you need a hypothesis-driven method to identify relevant questions.
• This creates new opportunities for smaller players to become visible for valuable questions – if they work strategically.

Is SEO dead?

Not according to the data. The claim that "SEO is dying" resurfaces at regular intervals. But every time, reality has shown the same thing: SEO does not die – the ways of working change.

The claim tends to surface with new SERP features (e.g. featured snippets, AI Overviews or other changes in the search results) or bigger technological leaps in search and in search engine algorithms. The problem is that many of these conclusions are based on anecdotal evidence or limited data sets, rather than broad and representative analysis.

As SEO best practice keeps evolving, there will always be studies on shaky foundations that point to declines and are misread as signs that SEO as a channel is about to disappear. Historically, these readings have been proven wrong time and again.

Graphite, a leading US SEO/AEO agency, worked with Similarweb (a major digital analytics platform) to analyse 40,000+ large US websites, comparing periods over time. The conclusion was clear:

• Organic SEO traffic is down -2.5% YoY – a modest decline in context.
• AI Overviews can lower CTR by -35% when they are present, but they appear about ~30% of the time according to the same analysis.
• The idea that "Google steals clicks for ads" is exaggerated; according to Graphite's data, organic is still much bigger than ad clicks.

Translation:
SEO has not "died". But clicks are shifting (and attributing conversions to their source is getting harder) as more questions are answered directly in the interface. So KPIs such as "sessions" mean less on their own, and you have to work towards visibility where decisions are made, not only where the clicks happen.

What has actually changed?

1) From "search phrase" to conversation (and more steps before the click)

• AI search makes it easier to ask follow-up questions, set the context and ask for recommendations – taking what has long been called "long-tail keywords" a step further. As a result, the research journey is often longer but more qualified once someone does click through.
• Webflow, one of our two preferred platforms for websites, reports a 6 times higher conversion rate on traffic from LLMs than on traffic from traditional search engines.
• AEO currently involves far fewer searches and therefore drives much less traffic, but that traffic is much further along in the decision-making process.

2) CTR becomes more volatile when AI answers appear

When AI Overviews appear, the CTR on organic results can drop (Graphite reports -35%). But they do not appear everywhere, and they show up more often on informational searches than on searches with commercial intent. And the trend looks set to accelerate: according to Ahrefs, AI Overviews correlate with a 58% lower average click-through rate (CTR) for the top-ranking organic result – up from 34.5% in an earlier measurement.

3) "Good SEO" is also "good AEO" – but with new additions

The things that make you visible in Google's organic search results are often the same things that make you appear in AI-generated answers. That is why classic SEO principles – such as schema markup, semantic heading structure, internal linking and technical structure – matter at least as much in AEO.

With one important caveat: not all technical measures carry the same weight when it comes to AI citations. When Ahrefs added schema markup to 1,885 pages, the AI citations barely moved. Treat schema as basic hygiene for crawlability and machine readability – not as the lever that gets you cited. The real levers are semantic structure, citable content and mentions.

I recently pointed out to a new client that their semantic heading structure was wrong throughout and needed fixing. They asked whether this was really necessary, since their previous web agency had said something along the lines of:
"Google and AI are so smart today that you don't need that any more."

If your web agency says that – contact us.

Off-site content: AEO makes brand mentions even more important and can lead to "Earned AEO"

Brand mentions are becoming increasingly important – and recent data suggests they carry more weight than many traditional signals. In Ahrefs' analysis of 75,000 brands, brand mentions correlated 0.664 with visibility in Google's AI Overviews – stronger than both backlinks (0.218) and domain authority (0.326). A follow-up (May 2026) points to mentions on YouTube as the strongest single factor (~0.737).
Make sure you are visible on platforms such as Reddit and YouTube, and through influencers, affiliates or partners. AI answers are influenced by what happens outside your website.

You can read more about Earned AEO in our glossary.

AEO creates new conditions for competition

In classic SEO, large companies have long had a clear advantage through high domain authority, strong link profiles and large content teams. As search becomes more long-tail and more conversational, the playing field changes.

The data shows that the playing field really is partly new: according to Ahrefs, 28.3% of ChatGPT's most-cited pages have no organic visibility in Google at all. In other words, they get cited even though they do not rank – a partly separate discovery layer where whoever gives the clearest answer can gain ground without being the biggest.

Thousands of narrower, deeper and more business-relevant search intents are emerging. That does not necessarily favour the biggest players – but those who are:

• quick on their feet
• specialised
• strategic in their content and add unique insights

The value of generic, broad content is falling (thankfully), while niche content that solves a concrete problem better than anyone else gets more traction.

"In our work at SEOS Design we can clearly see that companies that succeed in AEO do not produce more content – they produce more relevant content for a narrower context."

The same reasoning comes up in an episode of Lenny's Podcast, where Ethan Smith (CEO of Graphite) describes how AEO in practice rewards clear, relevant answers in the right context – rather than high domain authority.

Technology + strategy + good processes = a real competitive advantage

An important shift is happening here.

With modern technology and smarter ways of working, it is now possible to produce and maintain high-quality, AEO-ready content without huge resources. What used to require large content teams and long lead times can now be done by smaller, more focused organisations – provided they work the right way.

The key is not more content, but better processes.

That means:

• a clear method for identifying the right questions
• automation where it speeds things up
• deliberate, manual input where human insight is needed

This is crucial. Without human involvement you risk simply adding to the pile of derivative, generic content – which neither people nor AI systems reward.

With tools, automation and AI – for example through our sister brand AI Partner – smaller, innovative companies can:

• identify extremely valuable long-tail questions
• create structured answers written for both people and AI systems
• work strategically with content at a pace that used to be possible only for large organisations

In an AEO context, relevance, clarity, new insights and deep expertise often count for more than sheer domain strength. The result is a new competitive landscape in which even smaller companies can be cited on business-critical questions – if they work the right way.

It is the beginning of a more dynamic, fairer and considerably more strategically demanding era for SEO.

AEO makes the tail longer: from 50 keywords to 1,000 "educated guesses"

The most important shift in practice is not that "everyone stops googling". It is that the long tail gets much longer:

• More niche questions
• Longer prompts
• Deeper follow-up questions
• More context ("for our industry", "for our team", "with our requirements")

"Our experience of AEO work, in Sweden and internationally, is that AEO works best when hypotheses about the target audience's questions are tested and iterated systematically – not when content is created from assumptions without a method."

Why this requires a new method

In classic SEO you usually have a clear signal: search volume.
Because the questions people ask ChatGPT and similar services are not public, you need to make educated guesses: informed assumptions based on the best proxy data available.

That is not "guessing" – it is hypothesis-driven work.

How do you identify AEO questions?

We recommend three "proxies":

1. Traditional SEO data (Google as a proxy): Questions with search volume in Google are still an excellent indicator of what the market wants to know. A rule of thumb, based on the total usage of ChatGPT and Google, is that for every 20 Google searches there is one question in ChatGPT.
2. Social listening (Reddit, forums, Quora): If people ask the question on forums, they are probably asking AI too.
3. Internal business data (sales, support, onboarding): This is often a badly underrated source – and the one most closely tailored to your business, your customers and your market.

How do you write for chunking?

AI models do not read a page linearly. They pick out standalone sections ("passages"). That is why the content needs to be modular and self-contained.

Recommended structure for each section:

1. Direct answer first
   Answer the main question in the first sentence. It should make sense entirely on its own.
   Example:
   "AEO does not replace SEO – AEO builds on SEO and brand signals."
2. Context and method
   Explain why and how in 2–3 short paragraphs. That builds understanding and authority.
3. Evidence, examples and data
   Original data beats recycled content. If you have no data of your own, use clear examples and well-known sources – but be specific.
4. Standalone sections
   Every H2/H3 section should be readable and citable without any other context from the page.
5. Conversational language
   Phrase headings as questions where possible and where it makes sense. Studies by Graphite/Webflow show that question-based headings improve the match with AI intent and the likelihood of being cited.

Does AEO convert well?

Yes – but for now it is about quality, not quantity.

In the Webflow/Graphite case:

• Signup conversion from LLMs is reported as 24%, compared with 4% for non-brand SEO
• They also conclude that AI chat users can have a ~6× higher conversion rate than non-brand organic search traffic, thanks to the long tail and the conversational format.

This needs putting in perspective: LLM signups rose from 4% to 8% of their total signups over the period – a share that is likely to rise markedly over time.

Semrush reports that AI search visitors are 4.4× more valuable than traditional organic visitors, based on conversion rate. Source

At SEOS Design, we are convinced that the value of being visible in AI models will only increase.

How do you measure the results of AEO work?

As AI answers take up more space, value becomes harder to attribute in your analytics, and more value – or connection to your brand – is created before anyone clicks.

To measure real business impact, attribution models become crucial. Because many interactions with AI answers involve no direct click, simple questions such as "how did you hear about us?" at the point of conversion play a bigger role, especially for B2B and products with long buying cycles. Traditional click-based "last-touch" attribution will still exist, but it will probably underestimate how much impact AI answers actually have on the business. Source

AEO checklist

Here is a practical checklist that works for organisations with limited time and resources:

1. Pick 5–10 "money topics" – for example related to integration, use cases, comparisons, pricing
2. Build a query fan-out: ~10–20 sub-questions per topic (ideally including objections)
3. Write for chunking: the question in the heading, the answer in the first sentence, the evidence after
4. Make "educated guesses" using as many proxies as possible: (SEO data + questions on forums + sales/support data)
5. Build an entity profile (consistent: name, offering, case studies, expertise)
6. Earn mentions: partner pages, communities, podcasts, affiliates
7. Optimise the pages that convert (not just the blog) – for example, add an FAQ section to your product page
8. Enhance existing content: find content gaps on existing pages and fill them
9. Update content (signals freshness + lets you add new data and questions) – ideally once a year. SEOS Design has established processes for content updates that automate large parts of the work.
10. Connect to your product/solution in as many sections as possible, as long as it feels natural.
11. Do not over-optimise technical SEO: once you have a good base, you can focus on new content, content gaps and mentions. Technical SEO is important, but in 9 cases out of 10 it is better to focus on good content once a solid technical base is in place.

Bonus for those with more time:

1. Broaden your content strategy: e.g. videos, podcasts
2. Citation optimisation: identify the most-cited URLs for your topic and work to get mentioned there
3. Experiment! Test, evaluate, reproduce – but watch out for false causation

Want us to build a 90-day plan for SEO + AEO? We combine SEO with dedicated work as an AI SEO agency in one joined-up service. Contact us for free advice.