Journal
Is SEO dead now with AI? What has changed and what has not
AI answers are built from websites, so search optimisation still matters. What it rewards, and how you measure it, is shifting.
The short answer
No. SEO is not dead, but it is changing shape. Google's AI Overviews and assistants such as ChatGPT and Gemini still build their answers from websites they can find and read, so crawlable pages, a clear identity and genuinely helpful content matter as much as before. What has changed is how often people click, and how success should be measured.
Key points
- AI answers draw on the same open web that search engines index, so a site that cannot be found or understood is missing from both.
- Fewer people click through on simple questions, which makes rankings alone a weaker measure of success.
- The fundamentals still hold: pages that can be crawled, a business that is clearly identified, and content written to help people.
- Thin pages written only to catch a keyword are among the first to lose value.
- Nobody outside the AI companies fully knows how sources are chosen, so treat confident promises with care.
Why are so many people asking whether SEO is dead?
Because the results page looks different from the one most business owners learned about. Search for a simple question on Google and there is often an AI Overview at the top, a written answer assembled from several websites, sitting above the familiar list of links. Google has also added AI Mode, a conversational way of searching. At the same time, many people now ask ChatGPT, Gemini, Perplexity or Microsoft Copilot for advice they would once have typed into a search box.
When the answer appears on the screen, fewer people need to visit a website to get it. That pattern, often called zero-click search, existed long before AI (think of weather boxes and featured snippets), but AI answers extend it to far more questions. If your traffic reports show more impressions and fewer clicks, it is natural to wonder whether the whole discipline has stopped working.
It is worth remembering that SEO has been declared dead many times: when mobile search arrived, when social media took off, when voice assistants were expected to replace typing. Each time, the channel changed and the underlying job stayed. That job is making sure the right people can find a business, understand what it offers and trust it enough to get in touch. The question now is not whether that job exists, but where it happens.
Where do AI answers actually get their information?
Mostly from websites. When an AI assistant answers a question about a product, a service or a type of business, it usually combines two things: what the underlying model learned during training, and pages it retrieves from the web at the moment you ask. The second part is the one businesses can influence, and it depends on a search index.
Google is direct about this. Its documentation on AI features says that AI Overviews and AI Mode draw on pages in Google's search index, and that there are no extra requirements or special optimisations needed to appear in them beyond the normal requirements for appearing in Search. In other words, the route into Google's AI answers runs through ordinary SEO. OpenAI publishes the crawlers it uses, including OAI-SearchBot, which it describes as the crawler that surfaces websites in ChatGPT's search results. Microsoft's documentation says Copilot sends search queries to Bing when it needs web information to ground an answer.
So the picture is not "search versus AI". It is one open web feeding several surfaces. A page that cannot be crawled, or that does not say clearly what a business does, is weak raw material for every one of them. That is the strongest argument that SEO is alive: the AI systems people worry about depend on the very things SEO has always looked after.
What has genuinely changed for businesses?
Quite a lot, and pretending otherwise does not help anyone. The changes are less about the technical rules and more about how people search and what a visit is worth.
- Answers come before links. For simple informational questions, the reader may never need to click, so a ranking no longer guarantees attention in the way it once did.
- Questions are longer and more conversational. People describe their situation in full sentences and ask follow-up questions, which favours pages that answer a specific need rather than repeat a phrase.
- Being named matters as well as being linked. An assistant may recommend a business by name without sending a click, which still shapes who gets the enquiry.
- Several businesses can appear in one answer. The comparison happens inside the answer, before anyone visits a website.
- Measurement is harder. Google has only recently started showing separate figures for its AI features in Search Console, with limited detail, and most assistants share very little data with site owners.
What has not changed, even with AI answers?
The foundations are the same ones good SEO has relied on for years. First, pages must be crawlable and indexable. If a robots.txt rule blocks a crawler, if important text only appears after heavy scripts run, or if a page is marked noindex by mistake, neither a search engine nor an AI system can use it. This is unglamorous work, and it is still where many visibility problems start.
Second, the business must be a clear entity. Search engines and AI systems try to work out who an organisation is: its name, what it does, where it operates, and which profiles and listings belong to it. When the name is written three different ways, or the services described on the homepage do not match the ones on directory listings, the system has less confidence and may choose a competitor whose details agree. Consistent naming, a proper About page and structured data that matches the visible content all help.
Third, the content must be genuinely helpful. Google's guidance on helpful, reliable, people-first content asks whether a page offers original information, whether it would leave a reader feeling they had learned enough, and whether it shows real expertise. Those questions read as if they were written for AI answers, even though they predate them. Content that makes specific, accurate claims is easier to cite than content that says what every competitor says.
Which kinds of SEO are losing their value?
Some tactics are clearly weaker now, and it is better to say so plainly. The first is the generic article written to catch a keyword: "What is a dental implant?" or "Ten tips for choosing an accountant", written in the same way as hundreds of other pages. AI answers can summarise those topics without sending anyone to any of them, so the traffic these pages once earned is shrinking.
The second is mass-produced pages that differ only by a word or two. Search engines have long tried to filter these out, and AI systems have little reason to cite a page that adds nothing. The third is reporting that only tracks keyword positions. A position can look healthy while clicks fall, because an AI Overview now answers the question above it.
What holds up better is content tied to real decisions: how a service works, who it suits, what it involves, what it does not cover, and how to choose between options. Take a physiotherapy clinic. A general article about back pain competes with every health site on the internet. A clear page explaining its assessment process, the conditions it treats and how to book gives both a search engine and an AI assistant something specific to point to when someone asks for help nearby.
What should a business do about SEO now?
Start with the basics that affect every surface at once, then measure in a way that reflects how people now search. None of this requires special tricks, and you should be wary of anyone who says it does.
- Check that your important pages can be crawled by Googlebot, Bingbot and the search crawlers AI companies publish, and that nothing in robots.txt or your security settings blocks them by accident.
- Give each main service its own clear page that says what it is, who it is for, where you offer it and how it works.
- Use one consistent business name and description across your website, Google Business Profile (if you have a location) and the directories that matter in your sector.
- Add structured data that matches what the page visibly says, including markup that describes your business and links to its official profiles.
- Earn mentions from real, relevant websites: industry bodies, partners, local or trade press, and open, helpful answers in communities where your customers ask questions.
- Measure enquiries and sales from organic search, not only rankings, and keep a fixed list of questions you ask AI assistants each month to see whether you are mentioned.
What do we honestly not know yet?
A fair amount. The AI companies publish guidance on crawling and access, but they do not publish exactly how their systems choose which sources to cite or which businesses to recommend. Answers also vary: ask the same question twice, or from a different account or country, and the names can change. Anyone who claims to know the precise formula is guessing, or selling.
We also do not yet know where click behaviour will settle. Google has said that clicks from results pages with AI Overviews tend to be higher quality, with people more likely to spend time on the site they visit. That may be true, but it is Google's own observation and hard to check from outside, and experiences differ widely between sectors. A local service business and a news publisher will feel these changes very differently.
What does seem durable is the direction. Every major AI search surface relies on finding and reading websites, and every one benefits from sources that are clear, consistent and useful. Investing in those qualities is a sensible bet whichever way the interfaces develop. Investing in shortcuts aimed at one system's current behaviour is not.
There is also a practical reason not to wait for certainty. The work that helps AI answers, such as fixing crawl problems, clarifying service pages and earning genuine mentions, also helps ordinary search results, Google Business Profile and the people who land on your site directly. If AI search grows faster than expected, you are ready. If it grows more slowly, the same work has still improved how customers find and judge you.
So is SEO still worth paying for?
Yes, if it is the right kind of SEO. Paying for monthly reports that list problems nobody fixes was poor value before AI and is poorer value now. The work that matters is the work that changes your website: fixing what blocks crawlers, clarifying service pages, tidying how your business is described, and publishing content that answers real customer questions in detail.
That is how we approach it at Deylaxy. We make the SEO changes on your website for you rather than sending a list, nothing goes live without your approval unless you choose to let small changes go ahead, and every change is checked on the live site after it is made. We also check each week whether AI assistants and Google's AI answers mention your business, because rankings alone no longer tell the full story.
No one can promise that Google will rank a page or that an AI assistant will name a business; those systems make their own decisions. What a business can control is whether it gives them clear, accurate and helpful material to work with. That is still SEO. It simply has more places to show up than it used to.
Frequently asked questions
Should I move my SEO budget into AI search optimisation instead?
For most businesses, no, because they are largely the same work. Google says its AI Overviews and AI Mode rely on its normal search index, and other assistants also retrieve web pages. Crawlable pages, clear service content and a consistent business identity help in both. The sensible shift is in measurement: track whether AI answers mention you alongside rankings and enquiries, rather than funding two separate programmes.
Do AI Overviews take traffic away from websites?
They can reduce clicks for simple questions that the overview answers in full, and some site owners report more impressions with fewer clicks. The effect varies by sector and query type. Questions that involve choosing a provider, comparing options or booking a service are harder to settle inside a summary, so pages built around those decisions are likely to be less exposed than general information articles.
If I block AI crawlers, will my Google rankings be affected?
It depends which crawler you block. Google-Extended controls whether your content is used for some Google AI models, and Google says blocking it does not affect inclusion in Search. Blocking Googlebot, by contrast, stops Google crawling your pages for Search, and AI Overviews rely on the same crawling. Blocking search crawlers from AI companies, such as OpenAI's OAI-SearchBot, can stop your pages appearing in their search answers. Decide crawler by crawler.
How long does it take for SEO changes to show up in AI answers?
There is no fixed timeline. A change first has to be recrawled, which can take days or weeks, and assistants then decide for themselves which sources to use. Some assistants also rely partly on training data that updates far less often. The practical approach is to ask the same set of questions at regular intervals, record the answers, and look for a trend over several months.
Is it worth writing blog content now that AI can answer questions?
Yes, but the type of content matters. Generic explainers that repeat what every other site says are losing traffic, because AI answers summarise them easily. Content that adds something specific, such as how your service actually works, how to choose between options, or first-hand expertise, gives search engines and assistants a reason to cite you rather than a competitor.
Sources
- 1.AI features and your website (Google Search Central)
- 2.Top ways to ensure your content performs well in Google's AI experiences on Search (Google Search Central Blog)
- 3.Creating helpful, reliable, people-first content (Google Search Central)
- 4.List of Google's common crawlers (Google Search Central)
- 5.Overview of OpenAI crawlers (OpenAI)
See whether search engines and AI assistants can find, read and understand your website today, before deciding what to change.