AI Is Taking Customers And Your Clicks
AI is taking customers directly now, not just search traffic, and most founders are only braced for one of those two problems. A [UPrinting survey of 1,000 business owners run via Pollfish](https://www.uprinting.com/blog/ai-tools-replacing-services/) found 25% say they lost business in the last year because a customer used an AI tool instead of paying for their service. At the same time, [Ahrefs' analysis of 300,000 keywords](https://ahrefs.com/blog/ai-overviews-reduce-clicks-update) found that when an AI Overview appears on a results page, the top-ranking organic result loses 58% of its clicks, up sharply from the 34.5% drop the same research measured in April 2025. Those are two separate disruptions landing on the same business at the same time: customers skipping the purchase entirely, and the customers who do search never reaching the website at all. Ranking first used to be the win condition. Now a business can rank first and still lose the click, on top of losing the customer who never searched in the first place.
Front one: the customer who never pays you
The UPrinting/Pollfish survey is blunt about this: a quarter of business owners have already watched a customer solve their own problem with a free AI tool instead of hiring them. Separately, 65.5% of the same business owners worry that leaning on AI themselves makes their business feel less personal to the people they serve. That is the substitution front, and it was already covered in depth, which services are actually exposed to DIY-AI and which are not, in losing customers to AI, what to fix first. The short version: it hits commodity, finished-artifact work hardest, and barely touches judgment, accountability, and relationship work. This post is about the second front, the one that happens before a customer ever gets the chance to hire anyone.
Front two: the customer who searches but never clicks
Ahrefs tracked search results across December 2025 and found the damage is not limited to the top spot. The number-one organic result loses 58% of the clicks it would have gotten without an AI Overview present. Position two loses roughly half. Even a page ranking tenth, well off the first screen, loses close to 20% of its clicks. The pattern holds across the board: the AI Overview answers the question on the results page itself, and a shrinking share of searchers ever scroll past it to a website. A business that spent years and real budget earning the number-one ranking for its most valuable keyword can now watch well over half of that traffic evaporate without losing a single ranking position.
Rank #1 and lose 58% of the clicks that used to come with it. The ranking did not get worse. The prize for ranking did.
Why both fronts have the same root cause
It is tempting to treat customer substitution and the AI Overview click decline as two unrelated problems, one about services, one about search. They are the same problem wearing two outfits. In both cases, an AI system sits between the business and the customer, and decides what the customer sees or does next, whether that is a finished deliverable the customer can use directly, or a summarized answer on a results page the customer reads without clicking through. The businesses losing on both fronts are the ones whose value was never clearly distinct from a generic answer. The businesses that hold up are the ones an AI system, and the human reading its output, can point to as the actual source, not just one of many results saying the same thing.
What "being the cited source" requires that ranking never did
| Old SEO moat (ranking-first) | AEO moat (citation-first) |
|---|---|
| Keyword-optimized title and meta description | A clear, directly-answerable claim an AI model can lift and attribute |
| Backlink volume and domain authority | Specific, checkable facts, numbers, names, dates, that generic competitors cannot copy |
| Winning position #1 on the results page | Being the source an AI Overview or chatbot names when it answers the question |
| One strong landing page per keyword | Structured FAQ and schema markup an AI system can parse and quote directly |
| Traffic as the success metric | Citations and brand mentions inside AI answers as the success metric |
Ranking first is still worth doing, it is just no longer the finish line. The AI answer engine visibility audit and the B2B AI search readiness audit both walk through the mechanics of checking whether AI tools can actually find, parse, and cite a given business, which is now the harder and more valuable half of search visibility.
The fix: build for citation, not just for clicks
Three moves matter more than anything else here. First, replace generic claims with specific, attributable facts, exact numbers, named case outcomes, dated results, the kind of detail an AI model can quote with confidence and a generic competitor cannot simply repeat. Second, mark up content so it is easy to parse, not just easy to rank, FAQ schema, clear question-and-answer structure, and a direct one-sentence answer near the top of the page, the same structure that makes a page quotable in an AI Overview is the structure that makes a human skim it faster too. Third, treat the service itself the same way, the judgment-and-relationship work that AI cannot substitute needs to be named and priced as clearly as the commodity work that AI can, so a customer comparing a free AI tool against a business sees a real difference, not just a result that looks the same as everyone else's.
A two-front self-check before the next planning cycle
Pull the business's top five organic keywords and check two things for each one: does an AI Overview now appear above the organic results, and if an AI model were asked the same question directly, would it name this business as a source. If the answer to the first is yes and the second is no, that keyword's ranking is producing a fraction of the traffic it used to and the business is invisible to the growing share of searches that never reach a results page at all. Fixing that gap, alongside the service-level fix covered in losing customers to AI, is what AI Search (AEO/GEO) and a fractional CMO for startups are built to run as an ongoing system rather than a one-time audit.
FREQUENTLY ASKED
What percentage of business owners have lost customers directly to AI tools?
A UPrinting survey of 1,000 business owners run via Pollfish found 25%, roughly 1 in 4, say they lost business in the last year because a customer used an AI tool instead of paying for their service.
How much do AI Overviews reduce clicks to the top-ranking search result?
Ahrefs' analysis of 300,000 keywords from December 2025 found that when an AI Overview appears on a results page, the top-ranking organic result loses 58% of its clicks, up from a 34.5% drop the same research measured in April 2025. The decline extends down the page, position two loses roughly half its clicks and even a tenth-place result loses close to 20%.
What does it mean to be "the cited source" instead of just ranking first?
It means an AI system, whether a search Overview or a chatbot, names a business directly as the source of an answer, rather than the business simply holding a high position on a results page most searchers no longer scroll past. It requires specific, checkable facts and clear, parseable structure, not just keyword optimization and backlinks.
How can a small business respond to both the customer-substitution and AI Overview problems at once?
Both problems share a root cause, being indistinguishable from a generic answer. The fix is the same on both fronts: replace generic claims with specific, attributable facts an AI system can cite and a competitor cannot copy, and clearly separate and price the judgment-and-relationship work a customer cannot get from a free AI tool.
RELATED SERVICES
Delivered in 22+ markets worldwide.
Want this done for your business?
Free audit. No pitch. 24-hour turnaround.
PROOF THIS WORKS