SEO in 2026: is your CMS ready for Google AI Overviews and generative search?

In eighteen months, search has changed in nature. Google no longer simply ranks links: it writes answers. For marketing and digital leaders, the question is no longer "how do we get onto page one?" but "how do we stay visible when page one is an AI-generated answer?" And that battle is largely fought inside your CMS.

Clicks are drying up, and it's now measured

The trend is documented by independent sources. According to a Pew Research Center study based on the real-world searches of 900 US internet users in March 2025, roughly one Google search in five already displayed an AI-generated summary, and 58% of users had come across at least one within the month. That share has grown steadily since.

The impact on traffic is unambiguous. When an AI summary is present, users click through to a traditional link in only 8% of cases, compared with 15% when it is absent, a click-through rate cut almost in half. The links inside the summary itself are clicked in just 1% of visits. And 26% of users end their session right after reading the summary, versus 16% otherwise.

On the adoption side, Google announced in mid-2025 that AI Overviews were reaching more than two billion users a month. "Zero-click" search has become the dominant behaviour across a growing share of queries. Easy organic traffic is shrinking, and it is not coming back.

Being cited matters as much as being ranked

The second shift is more structural. Generative engines, Google's AI Overviews and AI Mode but also ChatGPT, Perplexity and Gemini, assemble an answer by pulling passages from several sources, then cite them. As a definitive guide from Search Engine Land on "generative engine optimization" explains, these engines assess fragments of content rather than whole pages, favour clear and immediately usable answers, and lean heavily on brand authority as corroborated elsewhere on the web: press mentions, reviews, authoritative sources.

Ranking in the organic top 10 no longer guarantees being the source the AI cites, and pages that were barely visible until now can become exactly that if they answer a question precisely. Visibility is shifting from ranking to citation, and "being citable" is becoming an editorial objective of its own. It depends directly on what your content management platform can do.

Where the CMS comes in

SEO has long been treated as a layer of optimisation applied on top of content. In the generative era, it is the architecture of the CMS itself that determines whether you are eligible to appear in AI answers. A CMS designed for a homepage and a blog is no longer enough. Four capabilities are now decisive.

• Structured data (schema.org). Article, Organization, FAQ, HowTo, Breadcrumb: this markup lets engines understand the meaning of your content, not just its words. A modern CMS has to generate and maintain it automatically, across thousands of pages, with no manual intervention.
• Headless and API-first architecture. To show up in AI Overviews as well as in ChatGPT, a mobile app or a voice assistant, your content has to be deliverable anywhere, cleanly structured and accessible via API. Content locked inside rigid HTML templates is increasingly hard to put to work, for you and for machines alike.
• Freshness and editorial governance. Generative engines reward up-to-date content. The ability to update, date and republish quickly therefore becomes a competitive advantage. Workflows, versioning, last-modified dates exposed properly: all of these are CMS capabilities.
• Performance and technical accessibility. Load speed, mobile rendering, but also access for legitimate AI crawlers (GPTBot, ClaudeBot, PerplexityBot): blocking them in your robots.txt amounts to making yourself invisible in generative search.

Five questions to ask yourself

Before investing in yet another SEO tactic, it pays to check the foundations. Here are the questions a marketing decision-maker should be putting to their technical team or their CMS vendor:

• Does our CMS automatically generate valid structured data, at scale?
• Is our content accessible via API so it can feed any channel, generative engines included?
• Can we update and redeploy content within minutes, with a modification date exposed?
• Are our pages fast, accessible and open to legitimate AI crawlers?
• Do we govern the brand as an entity (about pages, authors, consistency of mentions)?

If the answer to several of these questions is "no" or "I'm not sure", the problem isn't your keyword strategy: it's your platform.

The CMS becomes a visibility asset

Generative search doesn't kill SEO, it moves the work. Ranking gives way to citation, and page optimisation to structured content engineering. In this new landscape, your CMS becomes the infrastructure that decides whether your content will feed AI answers or stay invisible.

The organisations that treat their content platform as strategic infrastructure will be the ones Google, and the other generative engines, choose to cite. Search will keep changing. The question is whether your CMS keeps up.

FAQ

What are Google AI Overviews and how do they change SEO?

AI Overviews are AI-generated summaries Google displays above traditional results. They answer the query directly, so the user often has no reason to click through. Pew Research Center found that in March 2025 roughly one Google search in five already displayed one. SEO therefore shifts from winning a position to being the source the summary quotes.

Do AI Overviews reduce clicks to websites?

Yes, and it is now measured. According to Pew Research Center, when an AI summary appears users click a traditional link in 8% of cases, against 15% when it is absent. Links inside the summary itself are clicked in only 1% of visits, and 26% of users end their session right after reading the summary, against 16% otherwise.

What is generative engine optimization (GEO)?

Generative engine optimization, or GEO, is the practice of making content easy for generative engines to extract, quote and attribute. It works at fragment level rather than page level. Engines assess individual passages, favour clear and directly usable answers, and rely heavily on brand authority as corroborated elsewhere on the web.

How do you get your content cited by AI Overviews and ChatGPT?

Answer one precise question per section, in plain language, at the top of the passage rather than at the end. Expose valid structured data so engines understand the meaning of the content. Keep pages current and clearly dated. Build brand authority off-site through press mentions and reviews. And leave legitimate AI crawlers free to read the page.

Which CMS capabilities matter most for generative search?

Four, mainly: automatic, valid structured data at scale; headless, API-first delivery so content can feed any channel; editorial governance that allows fast updates with an exposed modification date; and technical performance that stays open to legitimate AI crawlers. A CMS built only for a homepage and a blog will limit your eligibility for AI answers.

Should you block AI crawlers such as GPTBot, ClaudeBot and PerplexityBot?

Only if you accept being invisible in generative search. Blocking them in your robots.txt removes your content from the pool those engines can quote and attribute. This is a business decision, to be taken engine by engine, rather than a setting left at its default.