Is Your Website Still Playing Twenty Questions?
AI Overviews are cutting clicks to top results. The next challenge is making websites as responsive as the search that brought visitors there.
We have spent years teaching people how to use websites. Click here. Choose a category. Narrow the results. Try another search term. Go back to the menu. Open three pages and compare them.
Then generative AI came along and showed people something very different. Ask a question. Get an answer. Ask a follow-up. Add some context. Change your mind. The experience adapts.
Once people get used to technology, understanding what they’re trying to accomplish, the limitations of the traditional website become much more obvious.
Someone can tell an AI assistant, “I’m looking for X, but I need something that works with Y, and I don’t want Z,” and get a response tailored to those requirements. Then they arrive on a company’s website and have to translate that same need into a navigation path or a few keywords in a search box.
The bigger issue isn’t search: It’s the experience waiting for people when they arrive.
Search is Showing Us What’s Coming
For years, we learned to adapt our behaviour to search technology. We reduced questions to keywords, scanned a page of links and pieced together an answer ourselves. Generative AI is reversing that relationship. Increasingly, we expect the technology to do more of the work.
We’re already seeing the impact in search. In February 2026, Ahrefs updated its research on Google’s AI Overviews and found that when an AI Overview appeared, click-through to the top-ranking organic result was 58% lower.
Think about that. A company can rank first and still get substantially less traffic because the answer the user wanted has already been assembled for them.
It’s tempting to see this primarily as an SEO problem, but I think the bigger implication is its impact on our expectations. Search is teaching people that they shouldn’t have to find the right page, understand how information is organised or know a company’s terminology. They should be able to express what they need and let the technology figure out the rest.
The question is what happens when those same people arrive on our digital properties.
The Experience Needs to Change, Too
Most digital experiences are still created before the visitor arrives. We build pages and navigation paths, define audience segments, and create personalisation rules based on journeys we anticipate in advance.
Traditional personalisation improved those experiences, but it still relies heavily on anticipating what someone might need. If someone belongs to this segment, show this content. If they took this action, present that offer.
Generative CMS introduces a different starting point: What is this person trying to accomplish right now?
A Generative CMS can use AI to interpret that intent and context and dynamically assemble an experience in response. That could mean surfacing a tailored answer, bringing together relevant product information or documentation, changing a call to action or assembling a broader experience around what that person needs.
Consider someone visiting a university website because they’re thinking about going back to school while working full-time. Today, they might have to navigate separately through programs, tuition, financial aid, schedules and admissions requirements separately. A generative experience could understand the underlying need behind the visit and bring together the relevant information.
The distinction is important. Personalisation chooses among experiences we’ve already created. Generative CMS opens the door to assembling an experience around the immediate need.
Of course, that raises an obvious question: If AI is assembling an experience dynamically, how does an organisation maintain control over what it says?
This is where retrieval-augmented generation, or RAG, becomes important. Rather than relying solely on what a language model already knows, RAG allows the system to retrieve information from approved enterprise sources (product information, documentation, pricing, policies, and other trusted content) and use it to inform the response.
Combined with defined presentation controls, it gives the experience room to adapt while keeping it grounded in information the organisation trusts.
That’s an important distinction between a generative experience and simply putting a general-purpose AI model on a website. The goal isn’t to give up control of the experience. It’s about changing what we control: providing trusted content, context, and guardrails while allowing the experience itself to become more responsive to the person in front of it.
Being Found is Only Half the Challenge
None of this means SEO stops mattering. But traditional search visibility and visibility within an AI-generated answer aren’t necessarily the same thing. Ahrefs analysed 4 million URLs cited by Google AI Overviews across 8,63,000 search results in research published in March 2026. Only 38% of those citations came from pages that also ranked in Google’s top 10 for the query.
For marketers, that expands the job in two directions. Our content has to work beyond the page on which we publish it. It needs to answer real questions clearly, contain current, trustworthy information, and be structured so that AI systems can understand and retrieve it.
But being part of the answer is only one side of the equation. If someone has just had a highly contextual interaction with an AI system and then comes to our site, sending them back into a conventional navigation tree feels increasingly like asking them to start over. That’s where content strategy and experience strategy begin to converge.
Designing for the Need, Not Just the Journey
For most of the web’s history, the page has been the basic unit of the digital experience. That shaped the way we worked as marketers: What pages do we need? Where should the content live? What’s in the navigation? What journey do we want someone to follow?
Those questions aren’t going away. But increasingly, the page may be an ingredient in the experience rather than the experience itself.
That changes what we need from our content, too. Content has to be structured, reusable, current and governed so it can be retrieved and assembled in contexts we didn’t individually design in advance. And that may be the bigger story behind what’s happening in search.
AI isn’t just changing how people find information. It’s changing their expectations for what happens when they ask for it. Search engines are already adapting to that reality.
The next challenge is making sure the experiences we own can adapt, too.
ALSO READ: Customer Context Layer is Becoming the New Role of the CDP
