There is a conversation I keep having with clients right now, and it always starts the same way. They look at their ad data, they see clicks falling, and they ask me the obvious question: should we stop paying for this?
I understand why they ask. But I think it is the wrong question, and I want to explain why.
Attribution is just a way of understanding how someone came to know your brand. The mistake most people make is assuming there is one moment, one click, one ad that made the sale. There almost never is.
Nobody clicks on an ad and buys a service on the spot. People buy from you when they have seen your name in several different places, over and over, until you are the name they trust. They see an article. They see you in a search result. They see your ad. They see you referenced somewhere else. By the time they fill in your form, you have already shown up in their world many times. The form fill is the last step, not the first.
That has always been true. What has changed is that the search itself now hides most of those steps from us.
Type a question into Google today and you often get an answer straight away, at the top of the page, in the AI Overview box. The person reads it. They ask a follow-up question. They get another answer. And they never click through to a single website.
This is what search looks like now. The information comes to the person, so the person stops travelling to find it. Your ad may still be on that page. Your website may still rank. But the click that used to prove it was working is disappearing, because there is no longer any reason to click.
This is not a small trend, and it is not slowing down. Semrush studied more than 600,000 keywords across ten industries over six months, from November 2025 to April 2026. They found that AI Overviews grew by an average of 71% on searches with commercial intent — exactly the searches where someone is comparing options and deciding who to buy from. AI Overviews and Google Ads now appear together on the same results page roughly twice as often as they did a year ago. In some sectors the shift is dramatic: in finance, commercial searches triggering an AI Overview grew by 231% in six months.
The pattern makes sense when you think about how people buy. The longer someone spends researching a decision, the more likely an AI Overview is to appear and summarise it for them. High-value, considered purchases — the kind most of my clients sell — are exactly where the AI box is taking over. So the more considered your product, the more this affects you.
Here is the part that matters. Less click data does not mean less business. It means the proof has moved.
I would be very nervous about any client switching off activity right now just because the clicks on it are falling. The clicks were never the point. The leads are the point. If your form fills are holding up or going up while your click data goes down, that is not a failing campaign — that is a campaign working in a world where clicks are no longer the receipt.
There is a hard truth for anyone running paid search here too. Google still does not tell advertisers which of their paid searches also showed an AI Overview. So the ground is shifting underneath paid budgets and the platform is not showing you where. That is all the more reason to stop judging each channel by its own clicks in isolation, and start looking at the whole picture: are the right leads still arriving?
If people are getting their answers inside the AI box, then the whole game becomes making sure your brand is one of the answers. And you do not get there by shouting louder on one channel. You get there by being present in the places the AI is reading from.
Look closely at any AI Overview and you will see it cites its sources — articles, publications, reference pages. That tells you exactly what to do. You need to be in those publications. You need your expertise written up, placed, and published where the models can find it. This is why I push so hard on media and articles for clients. It is the single biggest lever I have seen for showing up in AI search, and it does more than one job at once: third-party credibility, better SEO, and a real presence in the AI answers, all from the same piece of work.
The same logic explains two tactics that surprise people. The first is Reddit. Your customers may not all be on it, but the language models treat it as one of their largest training sources. Amplifying the content you already publish into the right, active communities puts your name into the data the AI learns from. The second is Wikipedia. It is a huge source for these models, and I have watched large, sophisticated companies quietly editing relevant pages for months — adding references to their own papers and content — for exactly this reason. Not to game anything, but to be present where the answers are built.
None of these are click-chasing tactics. They are all about being in the room when the AI writes the answer.
Stop grading your marketing on the click. That number is going to keep falling, and it is going to keep falling for everyone, because the search page itself is answering the question now.
Judge it on the leads instead. Build presence across many touchpoints — ads, an optimised website, media placements, and the sources the models read — so that whenever someone searches your space, your name is there. Then watch whether the right people keep filling in your form.
The brands that win the next few years will not be the ones with the most clicks. They will be the ones who understood, early, that the click was never the thing that mattered.
Sources: Semrush, "AI Overviews are expanding across commercial intent search" (July 2026), based on a six-month analysis of 600,000+ keywords across ten industries.