If you have heard some version of “Google is dead” or “your website won’t matter once people shop through AI,” the latest data points in a much less dramatic direction.
Customers are not replacing one path with another. They are adding another path.
Similarweb’s State of Ecommerce 2026 report says direct referrals from AI platforms have grown more than 200% over the past year, but still send much less traffic to ecommerce sites than search. More importantly, 89% of consumers who use AI while researching a purchase also use search.
That changes the practical question for a business.
The question is no longer, “Should I optimize for Google or AI?”
It is, “Can a customer discover me in one place, verify me in another, and still feel confident when they finally reach my website?”
That is a harder problem. It is also a much more useful way to think about what is changing.
AI is becoming a research layer, not a replacement for everything else
The interesting part of Similarweb’s data is not simply that AI usage is rising. We already know that.
It is how people behave after AI gets involved.
Similarweb says AI recommendations can materially influence purchase decisions even though direct AI referral traffic remains relatively small. Its 2026 report also found that journeys involving both AI and search convert better than AI-only journeys.
In other words, AI may shape the shortlist without owning the entire sale.
Consumers are less tidy than our marketing diagrams. They use whatever helps them feel confident.
The verification loop is the part businesses should pay attention to
Yext’s 2026 Consumer Search Behaviors Report surveyed 3,848 consumers globally about local search. Because Yext sells search and visibility software, its findings should be treated as company research rather than neutral public statistics. Still, the behavior it found lines up remarkably well with Similarweb’s ecommerce data.
Yext found that only 5% of AI users move directly from an AI answer to a purchase. After receiving an AI recommendation, 53% search Google or Bing, 49% visit the business’s website, 42% click the sources or citations provided by the AI, 28% check third-party reviews, and 20% check the business on social media.
That is not a funnel. It is a verification loop.
A customer might ask an AI assistant for three options, Google one of the names, scan the reviews, open the website, go back to the AI to compare something, and then finally buy.
From the customer’s perspective, this is perfectly reasonable. From a marketer’s perspective, it ruins the comforting idea that one channel deserves all the credit.
Here is the part I would actually pay attention to: AI can help you get considered, but the rest of your digital presence still has to survive inspection.
Your website may matter more, not less
Adobe’s retail data adds another useful piece.
In July 2026, Adobe says AI-referred traffic to U.S. retail sites was up 62% year over year. Those visitors converted 60% better than non-AI traffic, spent 59% more time on the site, and were 33% less likely to bounce.
That sounds like a strong case for AI visibility. It is.
But notice what happens at the end of the journey: the customer is still on the retailer’s website.
The AI did not make product pages, pricing, policies, trust signals, reviews, checkout, or site performance irrelevant. It delivered a more informed visitor to them.
Adobe also found a large machine-readability gap. In its July analysis, the average U.S. retail homepage visibility score was 61%, meaning substantial portions of many sites were not fully readable to AI systems.
So businesses now have two audiences to satisfy at once: people need to understand and trust the website, and machines need to understand enough of the website to recommend it accurately.
Fortunately, those goals are not opposites.
Clear product information, useful service pages, accurate pricing, structured facts, strong reviews, accessible text, sensible internal linking, and a fast site tend to help both.
The future may involve more sophisticated technical optimization, but the starting point is almost annoyingly familiar: make your business easy to understand.
Do not abandon SEO because AI traffic is growing
There is a real temptation to move budget and attention toward whatever has the steepest growth chart.
AI referral traffic can grow by triple digits partly because it started from a small base. Similarweb says generative-AI referrals still account for only a small share of retail ecommerce visits.
That does not make AI unimportant. It means the sensible strategy is additive rather than substitutive.
Keep the search presence that already captures demand. Improve how your site can be understood and cited by AI. Maintain the reviews and third-party proof customers use to verify recommendations. Make sure the website that receives the eventual visit actually closes the confidence gap.
This is less exciting than “SEO is dead.” It is also much closer to how customers appear to be behaving.
What I would do if this were my business
I would start with the customer journey we can actually observe.
First, check whether the facts that matter most to a buyer are explicit on the website: what you sell, who it is for, price or pricing logic where appropriate, availability, location, service area, returns, warranties, shipping, FAQs, and the reasons someone should trust you.
Second, inspect how those facts appear outside the website. Are Google listings accurate? Are recent reviews healthy? Are important third-party profiles current? Does social make the business look alive and credible?
Third, start measuring AI referrals separately in analytics instead of burying them inside generic referral traffic. The volume may still be small, but high-quality visitors are worth watching.
Fourth, keep investing in search where search is producing business. Do not dismantle a working acquisition channel because a newer one is growing quickly.
Finally, think less about “ranking in AI” as a mysterious new marketing discipline and more about giving recommendation systems good evidence to work with.
If an AI recommends you and the customer cannot verify what it said, you have not won anything yet.
The bigger shift
Discovery is fragmenting, but trust is consolidating around consistency.
Customers can begin in ChatGPT, Gemini, Google, Instagram, a review site, an app, or a marketplace. The business does not control the starting point anymore—if it ever really did.
What it can control is whether the same clear, credible reality shows up wherever the customer checks next.
AI is becoming an important part of the buying journey. The latest evidence does not show it replacing search, websites, reviews, and social.
It shows customers using more of them together.
Sources
Ready to scale your brand?
Let's talk about what growth looks like for your business.
Book a Free Strategy Call