Google is in trouble. It has a very profitable business built on traditional search. Now external pressure is forcing Google to release AI search. The problem is that AI search means fewer clicks, fewer websites that can survive to use search and ultimately less money for Google. At first glance, that seems like a problem for Google to solve. But the structural pressures that are reshaping the way Google searches are the same ones that are reshaping how you should shop.
It’s a trap
For nearly 30 years, Google and the open web have operated in simple exchange. Google crawled the web and returned traffic to content owners. Google should own the search engine and monetize the set of queries, while the website owner has the traffic to drive revenue for advertising, selling products, and creating more content. Google needed that content. That’s what drives search and keeps people coming back.
AI search is breaking exchanges. Searches without AI overview already have a 34% zero click rate. In AI overview, that rises to 43%, and in full AI mode it’s 93%. Less traffic means less ad revenue for publishers, which means less content, which means worse search. What Google needs to enable search is being eroded by AI search.
Publishers are well aware of their eroded situation. Cloudflare, which handles about 20% of the web’s traffic, now automatically blocks AI searches and evaluates each payment model. The odds ratio to referral tells the story: Google is 14:1, OpenAI is 1,700:1, Anthropic is 73,000:1.
For marketers, the saying goes hand in hand. As third-party publishers’ inventory shrinks and zero-click rates rise, brands that produce genuinely useful, well-written content have a structural advantage. Not just in terms of traditional SEO, but because AI engines cite sources when answering questions. If your content is not cited, you are not in the answer.
The problem of making money has no clean answer
Google’s current revenue model is extraordinarily good. A self-promoting live auction where advertisers compete for clicks, quality results stay relevant, and Google takes a margin without processing the content. Google monetizes the layer that sits between user intent and action.
AI search will involve a few clicks. Product comparisons that previously took three clicks now appear in one AI response. Fewer clicks, fewer sales, less income per inquiry. And that’s before AI agents start completing transactions automatically.
The question becomes how ads take to keep Google in the lead. Sponsored placement within AI answers – products that pay to appear in an answer rather than next to it. Cost per action, where advertisers pay when an AI completes a transaction on their behalf. Sponsored content, where Google’s AI recommends a product and the product pays for that recommendation. Each one is visible. None are proven. Everything requires more negotiation, more verification, more infrastructure than the auction itself.
The dynamic pricing and auctions that advertisers have spent years developing aren’t there yet in their new form, so the strategic question now is how to build familiarity with emerging formats before those models become established and early adopters have a set limit.
Google has navigated this before. AdWords did not appear in 1998. But that change added a revenue model on top of an existing product. This one requires a model change while the product changes underneath it, and whatever comes out won’t be clean or fast.
Product war is the most important part for marketers
Even if Google solves the content delivery problem and figures out how to monetize AI, it’s only relevant if users keep coming. That practice is no longer guaranteed.
OpenAI, Perplexity, Anthropic et al are all trying to get better. Unlike Google they have no revenue conflict in doing so – they can only serve the user at a faster pace. Google must weigh every product decision against its impact on the existing business. That makes them structurally slow at the same time as speed.
Ironically, Google knows all about this trade for better answers compared to existing income. The transformer architecture that powers ChatGPT, Gemini, Claude and every other major language model came from Google after that. They didn’t miss the AI wave. They built them. They couldn’t be the first to ride it in search without undermining their existing sources of income.
An important point
Google is not going away, and will find a version of this that works. But the change is real and is already affecting the way search works. Google is caught between moving fast enough to stay relevant and slow enough to protect the revenue that supports the changes. That tension won’t resolve quickly, meaning the search environment you plan to fight today looks very different than the one you’ll be using in three years. The question you have to live with now is what parts of your search strategy are built on assumptions that no longer hold.
