For twenty-odd years, the whole game rested on one quiet assumption: people type short, clipped phrases into a search box. "black leather boots." "garden bench uk." "birthday flowers next day." You bid on those phrases, you wrote ads around them, and the world made sense.
That assumption is now falling apart, and this week Google put hard numbers behind it.
People are searching in sentences now
Fresh data out of Google's AI Mode shows that the way people search has shifted from keywords to conversations. Instead of "garden bench uk", someone now types "what's a good weatherproof bench for a small patio that won't cost a fortune". It's longer, it's messier, and it's a proper question rather than a shopping fragment.
Google's own figures suggest this behaviour has already made most of the keyword strategies people built in 2025 look dated. That's a big claim, but it matches what I'm seeing in accounts: search terms reports are filling up with long, spoken-sounding queries that no sane person would ever have typed three years ago.
The good news is that Google Ads has been quietly getting ready for this. Broad match, smart bidding and the newer AI-driven matching are all built to understand intent rather than exact wording. If you're still clinging to a wall of exact-match keywords and a tight negative list, you're fighting the current. The accounts that win now are the ones that give the machine good signals and let it match the sprawl of real human questions to your products.
The clicks AI Overviews take are the good ones
Here's the part that should make every e-commerce owner sit up.
There's been a comforting story going round that when Google's AI Overviews answer a question at the top of the page and nobody clicks through, it's fine, because those were low-quality clicks anyway. People just wanting a quick fact, not customers.
A new study says that's wrong. When researchers compared the clicks that survive an AI Overview with the ones that vanish, they found no meaningful difference in bounce rate, time on site or how often people bounced straight back to search. In plain terms: the traffic AI summaries are absorbing is just as engaged and just as valuable as the traffic you keep. You're not losing tyre-kickers. You're losing real visitors.
I'm not telling you this to depress you. I'm telling you because it changes where your effort should go. If organic clicks are quietly leaking away at the top of the funnel, the paid channel and your product pages have to work harder to catch the people who do come through.
What this actually means for your ads
None of this is a reason to panic. It's a reason to shift a few habits.
Stop obsessing over exact keywords. Feed the algorithm intent instead. Clean conversion tracking, good audience signals and a sensible budget will do more for you now than another fifty exact-match variations. Let broad match and smart bidding reach the long, conversational queries your customers are actually using.
Make your product pages answer questions. If people are searching in full sentences, your pages need to speak that language. Not keyword-stuffed titles, but clear answers to the real questions buyers ask: what it's for, who it suits, why it's worth the money. That's what wins the click, and increasingly it's what gets you quoted inside AI answers too.
Watch your search terms report like a hawk. This is where the shift shows up first. The long, question-shaped queries appearing there are a live feed of how your customers now think. Mine them for the language people actually use, and pour that language back into your ads and your pages.
Don't assume falling organic traffic means falling demand. It might just mean an AI summary answered the question before your visitor reached you. The demand is still there. The route to it has moved.
The through-line
Search isn't dying. It's growing up. People are asking fuller, more natural questions, and the tools sitting between them and your website are getting cleverer about intercepting those questions.
The businesses that thrive through this aren't the ones with the biggest keyword lists. They're the ones who understand what their customers are really asking, structure their ads to let the machine find those people, and build product pages that answer like a helpful human would.
If you're not sure whether your account is set up for how people search now rather than how they searched in 2023, that's exactly the sort of thing I'd want to look at. Give me a call, and we'll go through it together.