The question people actually want answered is whether the results page still matters.
Short version: yes, less than it did, and unevenly. Traditional listings have not disappeared. They have been pushed down the page for a growing share of queries, and for some questions they are no longer where the answer is found.
Google AI Mode is the clearest expression of that shift, and it changes what visibility means without changing much about what earns it.
Google AI Mode: Navigating the Future of Search in 2026
What actually changed with Google AI Mode
The results page increasingly leads with a generated answer assembled from multiple sources, with links offered alongside rather than as the main event.
Three consequences follow.
- Informational queries lose clicks first. If someone wanted a definition, a conversion or a quick fact, the answer arrives, and the visit does not.
- Commercial queries hold up better. People comparing suppliers, checking prices or preparing to buy still tend to visit, because a generated summary rarely satisfies a decision involving money.
- Position matters differently. Being cited within an answer is a different objective from ranking first, and the two do not always coincide.
Content that works in this environment
The change rewards a particular kind of writing, and it is not a new kind.
- Answer first, elaborate after. Put the direct answer in the first two sentences under a heading, then add context and caveats. Systems extract the clearest early statement.
- Be specific enough to be usable. Dates, figures, named locations, defined processes. Vague claims cannot be quoted in an answer to anything.
- Cover the question properly. Thin pages that skim a topic are poor sources. Depth on a genuine question outperforms breadth across many.
- Say what you do not know. Content that qualifies its claims appears to be treated more favourably, and it is more honest, which is a rare alignment.
- Keep it current. Verifiable recency matters more than it used to, so an unmaintained page is a declining asset in a way it previously was not.
Technical work that supports Google AI Mode
Make content available without requiring JavaScript execution. Server-rendered or static output is readable by anything that requests the page.
Use semantic markup so structure carries meaning, and add structured data for your organisation, services, location and genuine frequently asked questions.
Keep pages fast. Where performance and environmental impact both apply, treat them as the same lever: reducing unnecessary data transfer and processing supports a faster experience while lowering estimated digital carbon, though environmental figures remain modelled rather than directly measured.
Check your robots.txt reflects a deliberate decision about AI crawlers, since many sites block them accidentally through a rule copied from elsewhere.
Looking beyond one company
The strategic risk is not this particular feature. It is depending on a single channel you do not control.
Diversifying is unglamorous and effective: an email list of people who opted in, genuine relationships with local and sector publications, a Google Business Profile kept current, and enough direct recognition that people search for you by name.
Branded search is the metric worth watching here. It moves when recognition is building; it is difficult for any algorithm change to take away, and it tends to be ignored in favour of rankings.
What nobody can tell you
How sources are weighted, how the balance between generated answers and listings will settle, or whether any particular page will be cited. These systems change without announcement, and the published detail is limited.
Anyone offering certainty about it is selling something. What can be said is that clear, accurate, well-structured, current content on a fast site has been the right answer through every previous shift, and there is no sign this one differs.
Our search engine optimisation work covers search and AI visibility as one brief.