The Long View / The Big Shift

Nobody searches anymore. They ask. Here’s who survives.

I started in marketing before the internet existed. I have watched the Yellow Pages give way to search engines, search engines give way to social feeds, and desktops give way to the phone in your pocket. Each of those shifts took about a decade, and each one made a lot of people rich and put a lot of people out of business.

What is happening now is faster than all of them, and it is a different kind of change. The earlier shifts changed where customers looked. This one changes whether they look at all.

Your customers have stopped searching. They have started asking.

The numbers nobody in marketing can ignore

I read the primary research behind each of these. They come from different companies using different methods, and none of them is a perfect census. Together they describe the same direction of travel.

  • Two out of three Google searches now end without a click. SparkToro, using Similarweb data, found that 68% of US Google searches in the first four months of 2026 ended with no click at all. The answer was on the results page, so nobody left.
  • When Google shows an AI answer, the top result loses more than half its clicks. Ahrefs compared 300,000 keywords between December 2023 and December 2025 and reported that the click-through rate of the number-one result fell by 58% when an AI Overview appeared above it.
  • Almost half of consumers now ask AI who to hire. In BrightLocal's 2026 survey, 45% of US consumers had used ChatGPT or another AI tool for local business recommendations, up from 6% a year earlier. AI became the third most popular source of recommendations in a single year.
  • ChatGPT alone has 900 million weekly users. OpenAI announced that figure in February 2026.

Read those together and the conclusion is simple. Being on page one is no longer the prize. Being in the answer is.

The funnel didn't change. It collapsed.

For twenty-five years, buying anything online followed the same path: you searched, you scrolled a list, you opened five tabs, you compared, you clicked, you decided. Every step in that path was a business. Search engines sold the top of the list. Directories sold the comparison. Review sites sold the trust. Ad networks sold the click.

The new path is two steps long. You ask. It answers. The comparing, the tab-opening and the scrolling now happen inside the AI, invisibly, in about two seconds.

How buying used to work, and how it works now.
The search eraThe answer era
The customerSearches with keywordsAsks a full question, in plain words
They seeTen links and some adsTwo or three names, with reasons
Comparison happensIn the customer's head, across tabsInside the AI, before they see anything
Your goalRank highBe named, and described correctly
What winsKeywords, links, ad budgetClear facts, consistent everywhere, vouched for by others

When the middle of a funnel collapses, the businesses that lived in the middle feel it first. So let's go through them one by one.

What happens to Yelp?

Yelp is the perfect case study, because it is living through this shift in public, in its quarterly reports.

In the second quarter of 2026, Yelp's total revenue grew just 1%, to $375.5 million. Its advertising revenue, the heart of the business, fell 3%, with weakness in restaurants and retail and fewer ad clicks. Meanwhile one small line in the report nearly doubled: "other revenue," up 98% to $33 million, driven by data licensing, subscriptions and acquisitions.

Then, in July 2026, Axios reported that Yelp is licensing its reviews, photos and business information to OpenAI, so that ChatGPT can show Yelp ratings and reviews inside its answers. Yelp's chief executive, Jeremy Stoppelman, put it bluntly: "If you want to answer local queries, you really need Yelp."

That tells you exactly where this is going. Yelp is not dying. It is changing what it is. It is moving from being a destination people visit to being a source that AI reads. Fewer people will open the Yelp app to choose a restaurant. Far more people will see Yelp's reviews without ever knowing they came from Yelp.

The same will happen to every directory and listing site. The ones with large, trusted, well-structured data become ingredients in AI answers. The ones that were only a list of ads with a search box are in real trouble.

Are reviews still relevant? More than ever. Just not the way you think.

This is the part most business owners get wrong. They assume that if AI answers the question, nobody reads reviews anymore. The opposite is true.

BrightLocal's 2026 survey found that 97% of consumers read reviews for local businesses, and the share who always read them jumped from 29% to 41% in a single year. And now there is a second reader: the AI itself. When someone asks for "the best family dentist near me for nervous kids," the AI is reading thousands of reviews and summarising them in a sentence.

That changes what makes a review valuable:

  • Specific beats glowing. "Dr. Lee was patient with my anxious six-year-old" teaches an AI who you are good for. "Great service!!!" teaches it nothing.
  • Recent beats old. A steady flow of new reviews tells the AI you are still open and still good.
  • Everywhere beats one place. Consumers in the BrightLocal survey used an average of six review sites, and Google's share fell from 83% to 71% while Apple Maps nearly doubled to 27%. AI systems draw on many sources. A business with reviews on only one platform has a single point of failure.
  • Fake reviews are now poison. They were always dishonest. Now they are also dangerous, because AI is trained to notice patterns, and a pattern of suspicious praise can quietly remove you from the answer.

Reviews are no longer just social proof for the next customer. They are the raw material AI uses to decide whether you exist, and what you are good at.

What happens to the companies that built apps for Shopify?

On December 10, 2025, Shopify announced that merchants can ask Sidekick, its built-in AI assistant, to create custom apps for their store. You describe what you need, and Sidekick writes the code, builds the screen and connects it to your store's data. The tool started with merchants on the Grow, Plus and Enterprise plans.

Now think about what that means for the Shopify App Store, which held 17,891 apps from 11,352 developers in May 2026 according to AppJubilee's monitoring. A large share of those apps do one simple thing: add a banner, calculate a shipping estimate, show a size chart, export some orders. Thousands of small businesses were built on charging $9.99 a month for that one simple thing.

When a merchant can type "make me a size chart that changes by product type" and get it in minutes, the one-trick app becomes a feature, and features don't charge rent.

This won't wipe out the whole app economy. Here is how I expect it to sort itself out:

  • In danger: single-purpose apps whose whole value is a bit of code anyone can now describe in a sentence.
  • Safe for now: apps that hold something the merchant can't generate, such as a network of other stores, years of data, payment and tax compliance, shipping carrier contracts, or a real service with people behind it.
  • The new winners: developers who stop selling code and start selling outcomes, like "we recover abandoned carts and you pay us a share of what we recover."

I build software for Shopify stores myself. The lesson I take from this is that the value has moved from writing the code to knowing what to build and why. That is a strategy question, not a programming one.

Shopping inside the chat: the experiment that taught everyone a lesson

In 2025, OpenAI launched Instant Checkout so people could buy products without leaving ChatGPT. Etsy merchants came first, then Walmart, PayPal, Target and Instacart joined. For a few months it looked like the end of the online store as we knew it.

Then reality arrived. Merchant integration was slow and expensive, people browsed far more than they bought, and in March 2026 OpenAI changed course: ChatGPT now sends shoppers to the retailer's own app or website to complete the purchase.

That reversal holds the most practical lesson in this whole article. Discovery has moved to AI. The purchase, and the relationship, still happen on your property. Your website is not obsolete. It has a new job: to be the place an AI confidently sends people to, and the place that closes the sale when they arrive.

The pattern behind all of it

Step back and every one of these stories is the same story.

The businesses that sold attention are losing. The businesses that own the truth are winning.

Who loses value and who gains it as AI becomes the front door.
Losing valueGaining value
Directories that were mainly ads and a search boxSources with deep, trusted, structured data (reviews, inventory, credentials)
Ranking on page one for broad keywordsBeing clearly described, consistently, everywhere
Simple single-purpose softwareKnowing what to build, and owning the data it runs on
Buying more clicksEarning mentions, reviews and citations
Marketing that is louderMarketing that is clearer and provable

AI is, at heart, a machine for reading what the world says about you and repeating the most trustworthy version. It doesn't care how much you spent on ads last month. It cares whether the facts about you are clear, consistent and confirmed by people other than you.

Five predictions for the next two years

These are my judgment, not data. I'll revisit them publicly in a year.

  1. "Search engine optimization" splits in two. Classic SEO keeps the website technically healthy; a second discipline, often called GEO, AEO or LLMO, takes over the work of getting named in AI answers. Most businesses will need both.
  2. Review platforms become data companies. Yelp's licensing deal is the template. Expect more review and directory sites to earn more from AI companies than from small-business advertising.
  3. Thousands of small software tools disappear into the platforms. Shopify's Sidekick is the first; every major platform will let customers generate their own simple tools.
  4. Advertising moves inside the answer. Paid placement inside AI conversations will grow, clearly labelled. But the organic answer, the one nobody can buy, will carry more trust.
  5. The small business with the clearest story beats the big business with the biggest budget. For the first time in a long time, clarity scales better than spend.

What to do on Monday morning

You don't need to understand all of this to act on it. Seven things, in order:

  1. Ask the AI about yourself. Ask ChatGPT, Gemini and Perplexity the question your best customer would ask, without using your name. Are you in the answer? Who is, and why?
  2. Fix your facts. Make sure your business name, address, phone, hours and services are exactly the same on your website, Google, Apple Maps, Yelp, Facebook and your industry directories.
  3. Write one clear sentence that says what you do, for whom and where, and put it at the top of your home page instead of a slogan.
  4. Ask for specific reviews on more than one platform. Ask customers to mention what you did for them, not just how many stars.
  5. Answer real questions on your website, in plain language, one question per section. That is what AI quotes.
  6. Stop paying for apps and tools you could now replace with a few sentences to an AI assistant, and put the savings into the steps above.
  7. Measure it every month. Ask the same questions again and write down who gets named. It is the only scoreboard that matters now.

I have spent forty years watching technologies change the way people find businesses. Every time, the same thing was true: the tools changed, but the businesses that won were the ones that were genuinely clear about who they were and genuinely good at it. That hasn't changed. What's changed is that, for the first time, a machine is checking.

Want to know what AI says about your business today? The free scan shows which AI systems know you, and what they say.  Run the free audit →

Common questions

How is AI changing marketing in 2026?
Customers increasingly ask AI assistants such as ChatGPT, Gemini and Google's AI Overviews for recommendations instead of searching and comparing links. In 2026, about two-thirds of US Google searches ended without a click, and nearly half of consumers had used AI for local business recommendations. Marketing is shifting from winning clicks to being named, accurately, in the answer.

Will Yelp survive AI search?
Very likely, but in a different form. Yelp's advertising revenue fell 3% in the second quarter of 2026 while its other revenue, including data licensing, nearly doubled, and in July 2026 it agreed to supply reviews and business information to ChatGPT. Yelp is becoming a source that AI reads rather than only a destination people visit.

Do online reviews still matter with AI?
Yes, more than before. 97% of consumers read local business reviews in 2026, and AI assistants now read and summarise reviews to decide which businesses to recommend. Specific, recent reviews spread across several platforms carry the most weight; fake reviews carry growing risk.

What happens to Shopify app developers now that merchants can build their own apps?
Since December 2025, Shopify merchants on certain plans can ask Sidekick to generate simple custom apps for their store. Single-purpose apps that do one simple task face the most pressure. Apps that hold data, networks, compliance or real services, and developers who sell outcomes instead of code, are best placed to survive.

Is SEO dead?
No. A healthy, crawlable website with useful content remains the foundation, and AI systems draw heavily on the open web. What has changed is the goal: beyond ranking, a business now needs to be clearly described, consistent across sources and vouched for by others, so AI systems can confidently recommend it.

What should a small business do first?
Ask the main AI assistants the question your best customer would ask and see whether you are named. Then make your business facts identical everywhere, state clearly what you do and for whom, collect specific reviews on several platforms, and re-check the same questions every month.

Sources

About the author. Siamak Kalhor is an AI consultant in Los Angeles with more than forty years of experience in marketing strategy, and the creator of the S.C.A.L.E. Framework™. He helps businesses become the answer AI assistants give. Get in touch, or connect on LinkedIn.