Consumers Are Skipping Google and Asking AI What to Buy

Google made an AI-generated answer the default response to every search query worldwide on July 10, retiring 27 years of ranked blue links. Four days later, marketers are already describing a customer who skips the search bar entirely and asks a chatbot what to buy instead.

Ankur Parmar, chief marketing officer at Mahindra Lifespace Developers Ltd, an Indian real estate developer, had already told the marketing trade publication Exchange4media that this was happening before Google made it the default. His read on the customer journey lines up with what just became true for everyone building a brand online.

Google Retires the Blue Link

Every search query now returns a real time answer written by Gemini 3.5 Flash, Google’s newest inference model, before it returns anything else. The tech outlet Tech Times, which first detailed the completed rollout, called it the most sweeping transformation of Google’s flagship product in the company’s 27 year history. The ranked list of ten blue links has not disappeared. It usually sits below the fold now, when it appears at all.

The retreat from links was already measurable well before July arrived, across four separate benchmarks.

Each benchmark measures something different, a click, a share, a ranking, but all four point toward the same collapse in traffic leaving Google for the open web.

The overhaul was never just about text answers, either. Days after Google’s I/O announcement in May, the company paired its Gemini app with Walmart, Shopify and Wayfair so shoppers can check out without leaving the chat. John Furner, Walmart’s incoming president and CEO, called the arrangement “the next great evolution in retail,” framing it as a shift from web and app search toward agent-led commerce.

A Real Estate Marketer Already Made the Switch

Parmar’s comments to Exchange4media read like a preview of Google’s July rollout rather than a reaction to it.

People are not Googling, they are talking to AI for what they need to buy.

Parmar told Exchange4media, adding a plainer line right after it: “AI cannot be avoided.”

He is not describing a fringe habit. HubSpot’s 2026 State of Marketing report found 44% of AI search users now call it their primary source for product discovery, ahead of traditional search at 31%, retailer websites at 9% and review sites at 6%. ChatGPT alone reported 900 million weekly active users by February, roughly double its count a year earlier.

Inside his own marketing team, Parmar says the shift already has a name, one he calls GEO. “The SEO which we used to do, now is GEO training,” he said, describing how the same keyword research that used to feed search rankings now feeds AI citations instead.

That does not mean the company website stops mattering. “The website remains the most authentic place to get product information in its entirety,” Parmar said, arguing the site becomes the anchor a shopper eventually checks even when a chatbot delivers the first answer.

He also let go of the tidy marketing funnel. Awareness, consideration and purchase no longer move in a straight line for a buyer weighing a multi year home purchase, he said. Buyers double back and return to the same question days later, so Mahindra Lifespaces tries to repeat the same accurate information at every step rather than chase one channel.

Authenticity Outranks Repetition

Parmar’s read on how an AI actually ranks a brand differs from classic search logic in one specific way. He described an engine that will skip a source it has seen mentioned constantly if a less frequent one carries more credibility, favoring an authentic reference over sheer repetition when it builds an answer.

• Generative Engine Optimization (GEO) – the practice of structuring content so AI systems such as ChatGPT, Gemini and Perplexity cite and recommend a brand inside a generated answer, instead of simply ranking its page on a results list.

That instinct lines up with what researchers have measured directly. A team from Princeton, Georgia Tech, the Allen Institute for AI and IIT Delhi coined the term Generative Engine Optimization in a paper presented at the KDD 2024 conference in Barcelona, and found that specific techniques, citing sources and adding statistics among them, could lift a brand’s visibility inside AI answers by as much as 40%.

A separate analysis of more than 25 million links by the media intelligence firm Muck Rack found 84% of AI citations trace back to earned media, coverage a brand did not pay for or publish itself, rather than a company’s own website.

A benchmark from the AI visibility tracker CiteLens found the split varies sharply by engine. Google’s AI Mode and Perplexity draw about 90% of their citations from pages already ranking in Google’s top 10 results, while ChatGPT pulled just 30% from that same pool, evidence that ranking well on Google guarantees far less inside a chatbot than it used to.

Who Wins When Nobody Clicks?

AI platforms and the retailers wired directly into them win first. A shopper who asks ChatGPT or Gemini to compare products can now finish the purchase without opening a browser tab, and OpenAI, Google, Amazon and Meta are each racing to own that final click before a rival does.

The commerce layer behind that shift is filling in fast, platform by platform.

• ChatGPT added shopping features in March that let users browse, compare and refine purchases conversationally, and Walmart is folding its own assistant, Sparky, directly into the chat window.
• Google Gemini runs a Universal Checkout Platform that lets shoppers pay with Google Pay without leaving the conversation, and the retailer Gap now sells directly inside it.
• Meta AI is quietly testing a shopping research carousel for a small group of US users, arriving last among the big three chatbots but backed by Facebook, Instagram and WhatsApp.
• Amazon’s Alexa+ supports conversational shopping through partners like Expedia and Yelp.
• Microsoft Copilot is testing a Checkout feature so a shopping chat never has to end at an outside link.

None of it guarantees a sale yet. Reviews that pit the assistants against each other for a real purchase still find gaps, with ChatGPT’s familiar chat layout and generous product links getting testers closest to checkout, while Gemini’s plainer default interface leaves more friction.

The Small Brands Nobody Cites

Being cited by an AI engine is not evenly distributed. A paper measuring brand visibility across AI search engines found that big, already authoritative names get cited almost no matter what they do next.

“The hard, unsolved problem belongs to everyone else: the SMEs, the D2C brands, the creators, the early-stage startups, and any brand without much of an online footprint,” the paper’s authors wrote.

The publishers whose work actually trains those answers are faring worse than most brands. HubSpot has reported losing between 70% and 80% of its organic search traffic. Chegg reported a 49% decline, and DMG Media documented drops as steep as 89% on some queries. NPR called the shift an “extinction-level event” for online news publishers.

Regulators and courts are starting to weigh in, too. Antitrust cases tied to the AI Overview rollout are moving through courts on two continents, and a German court has already ruled that Google’s AI generated answers count as the company’s own words, meaning it can be held liable when they are wrong.

Billboards Still Get a Budget Line

Parmar’s own budget undercuts any idea that AI anxiety means an all digital pivot.

“We do a lot of print. We do a lot of OOH. Digital is part of our plan too,” he said. “We don’t invest disproportionately in digital. It commensurates with its share.”

New channels get added deliberately, he says, without pulling budget away from the traditional media he credits with building the brand in the first place. Consumers have not abandoned billboards and newspapers just because they now open a chatbot first.

Google’s ten blue links have not vanished. They just are not the first thing said back anymore, and marketers like Parmar spent months rebuilding their playbook around that fact before Google ever made the change official.

Frequently Asked Questions

What Is Generative Engine Optimization (GEO)?

Generative Engine Optimization, or GEO, is the practice of structuring content so AI systems such as ChatGPT, Gemini and Perplexity cite and recommend a brand instead of just ranking its page. The term was coined by researchers from Princeton, Georgia Tech, the Allen Institute for AI and IIT Delhi in a paper presented at the KDD 2024 conference in Barcelona.

Did Google Remove Its Search Results Page?

No. The ranked list of links still exists, but it usually loads below the AI generated answer rather than above it. Google has added a Further Exploration section at the end of its AI answers with curated links, an attempt to send some clicks back to the sites its answers are built from.

Why Would an AI Recommend a Competitor Instead of My Brand?

Competitors displacing an existing citation cause the loss about 80% of the time, according to 2026 brand-citation tracking research from Erlin, rather than a technical error or an outdated page. AI engines weigh how often independent, credible sources mention a brand more heavily than how often the brand mentions itself.

Can You Actually Buy Something Inside ChatGPT or Gemini Right Now?

Yes, for a growing list of retailers. Walmart customers can complete instant checkout inside ChatGPT for nearly everything on its site except fresh food. Sephora has launched a ChatGPT app that folds in loyalty points and personalized recommendations, with in app checkout still to come.

Is Traditional SEO Dead?

Not yet. Traditional organic search still sent roughly 345 times more visits than ChatGPT, Gemini and Perplexity combined as of late 2025, according to Ahrefs. GEO builds on top of conventional SEO rather than replacing it, since most AI citations still overlap heavily with pages that already rank well on Google or Bing.