Skip to main content

What we found and fixed for ecommerce brands. Learn more about what we found and fixed for ecommerce brands

All resources

Guide16 min read

The Future of Search With AI Answers: What Changes in 2027

Search is not ending, but the answer is now a place where buyers learn which brands exist, and fewer of them click through afterwards. In a 2025 Pew Research Center study, users who saw an AI summary clicked a result link in 8% of visits, against 15% without one.

By Maksudur Rahman, FounderPublished Last reviewed

Key takeaways

  • AI answers now sit above or inside search, and the tools behind them report audiences of a billion people or more.
  • Clicks to websites fall when an AI summary shows, in three studies (one of them shows correlation only), and Google says the clicks that remain are better.
  • What has not changed: a page still has to be indexed, readable and useful, and Google says SEO best practices still apply.
  • By 2027 the direction is search agents that research and book, several AI tools instead of one, and measurement by appearance as well as rank.
  • A brand prepares by keeping the basics working, writing down buyer questions, publishing original facts, being mentioned by sources people trust, and measuring appearances.

What is changing in search?

Search is changing in five ways that matter to a brand, and each one has evidence behind it. The common thread is that for some searches an AI-written answer now sits above the list of links, and the studies below measure what that does to clicks and to which brands get named.

  • AI answers now sit above or inside the results.
  • Fewer clicks reach websites.
  • AI answers reach searches where people compare options.
  • Search is becoming a conversation.
  • Brands get named, linked, or both.

Google AI Overviews and Google AI Mode put a written answer in front of the buyer, and ChatGPT, Gemini and Perplexity are built around answers too. The audiences are large. In a November 2026 Think with Google article for advertisers, Google says it sees over 5 trillion searches a year (as of January 2025) and that AI Overviews has 2.5 billion monthly users (as of May 2026).

Google's May 2026 Search announcement says AI Mode has passed one billion monthly users, with queries more than doubling every quarter since launch. OpenAI, October 2026, says ChatGPT reaches 1.2 billion people each week, and Alphabet, Q3 2026, says the Gemini app is at one billion users.

Every figure except Perplexity is the vendor's own, measured over a different period, so do not add them together. The Perplexity row is a survey share, not a user count.

ToolFigureWho reports itPeriod
ChatGPT1.2 billion peopleOpenAIEach week, stated October 2026
Google AI Overviews2.5 billion usersGoogleMonthly, Google internal data, May 2026
Google AI ModeMore than 1 billion usersGoogleMonthly, stated May 2026
Gemini appOne billion usersAlphabetThird quarter 2026 remarks, no period given
PerplexityListed at 9% in a usage breakdownSemrush surveySurvey run in December 2025

Source: each vendor page linked in the table. The Perplexity figure is a share listed in a usage breakdown in a survey of 1,030 US consumers who had tried AI tools, not a user count.

Fewer clicks reach websites

When an AI summary appears, fewer people click a result link. The Pew Research Center, July 2025, tracked the browsing of 900 US adults in March 2025. Users who saw an AI summary clicked a traditional result link in 8% of visits. Users who did not see one clicked nearly twice as often, in 15% of visits. Clicking a link inside the summary itself happened in just 1% of visits to pages that had one.

Two other studies point the same way. Researchers at Carnegie Mellon University and the Indian School of Business ran a randomised field experiment, posted as a 2026 working paper on SSRN. They found that when AI Overviews appear, they reduce outbound organic clicks by 39.8% and increase zero-click searches by 34.5%, without affecting sponsored clicks or overall search frequency. It has not been peer reviewed as we read it.

An Ahrefs study of 300,000 keywords, February 2026, compared December 2023 with December 2025 and found that an AI Overview now correlates with a 58% lower average click-through rate for the top-ranking page. The rate is average desktop click-through, from aggregated Search Console data. Ahrefs says correlates, and so do we.

Google sees it differently. Google Search Central says that when people click from results pages with AI Overviews, the clicks are higher quality, meaning users are more likely to spend more time on the site. That is Google's own claim. The field experiment above found that engagement on downstream websites, for people who did click, was not different across groups. So the two sources disagree on what a click is worth, and the evidence for fewer clicks is stronger than the evidence for better ones.

Share of Google visits with a click on a traditional result link

Visits with an AI summary ended in a click on a result link in 8% of cases, against 15% for visits without one.

  • AI summary shown8%
  • No AI summary15%

Source: Pew Research Center, July 2025, browsing data from 900 US adults in March 2025.

AI answers reach searches where people compare options

AI Overviews are spreading into commercial-intent searches, where people compare options. In a Semrush study, July 2026, of over 600,000 keywords across 10 industries, covering November 2025 to April 2026, the share of commercial results pages with an AI Overview grew 71% in six months. Semrush defines commercial intent as someone in research mode, evaluating options, comparing products or looking for deals. The keywords come from Semrush's US desktop database.

Searches closest to buying went the other way. The same study found the appearance of AI Overviews for transactional-intent queries decreased by 5%, and says that when a person is ready to make a purchase there is less chance of an AI Overview appearing. It also found that Google Ads and AI Overviews now appear together on the same results page roughly twice as often as they did a year ago.

For an ecommerce store, this means a buyer comparing products can meet an AI answer before the product listings, which can shape which names the buyer sees first. A buyer who is ready to check out is less likely to meet one.

Search becomes a conversation

Search is moving from short keywords to longer questions. Google's CEO said on Alphabet's third quarter 2026 earnings call that queries in AI Mode are on average three times longer than traditional searches, and that one in five AI Mode searches globally use voice, images or documents.

A longer question also gets handled differently behind the scenes. Google Search Central says both AI Overviews and AI Mode may use a "query fan-out" technique, issuing multiple related searches across subtopics and data sources to develop a response.

This means one buyer question can become several searches the buyer never sees. A page can be pulled in for a sub-question it was not written to answer, so pages that cover the follow-up questions around a product have more ways to be used.

Brands get named, linked, or both

An AI answer can use your page without naming your brand, and it can name your brand without linking your page. In a Semrush study with Kevin Indig, June 2026, the team logged 3,981 domain appearances for 115 prompts run across 14 countries and four leading AI search engines. Almost 62% were what the study calls ghost citations: the platform used the page as a source link, but the brand name never appeared in the answer.

The tools also differ from each other. The same study found that Gemini mentions brands in 83.7% of appearances but generates a citation link only 21.4% of the time. ChatGPT does the opposite: it cites brands 87% of the time but mentions them in only 20.7% of answers. The sample is small, so read it as a pattern worth checking, not a law.

Being a source is not the same as being known. A brand wants both, and needs to track both, which is what we cover in how to measure AI search visibility.

How domain appearances in AI answers split between citing and naming

Most appearances were links with no brand name in the answer, and just 13.2% both named and linked.

  • Cited, brand not named61.7%
  • Named, not cited25.1%
  • Named and cited13.2%

Source: Semrush with Kevin Indig, June 2026, 3,981 domain appearances across 115 prompts.

What has not changed in search?

Three things still decide whether a page can be found and chosen. None of them is new, which is the good news for a brand that has already done the basics.

  • Pages still need to be found and readable.
  • Original, useful content still wins.
  • Buyers still compare and check.

Pages still need to be found and readable

A page that Google cannot index cannot be used by Google's AI features. Google Search Central says that to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet. It adds that there are no additional technical requirements.

Asked whether SEO still matters, Google's generative AI optimization guide, last updated in July 2026, answers "In short, yes!" The reason it gives is that its generative AI features on Google Search are rooted in its core Search ranking and quality systems.

Both statements are Google's, and both describe Google's features only. They are still the clearest vendor statement on the question, and they match what the rest of this post finds.

Original, useful content still wins

Google's guide says that creating content people find unique, compelling and useful will likely influence a website's presence in generative AI search in the long run more than any of the other suggestions in the guide.

For a store, unique and useful means real product facts: sizing, compatibility, what is in the box, delivery times and return terms, written for the question a buyer is asking. Rewritten manufacturer text gives an AI answer nothing it cannot already get elsewhere.

Buyers still compare and check

Shoppers who use AI still use everything else too. In a Semrush survey, December 2025, of 1,030 US shoppers who had tried AI tools, 77% said they use AI and traditional search together. The survey's conclusion is that AI determines which brands make the shortlist, but other channels still matter for purchase decisions. Only 20% said they use AI throughout the entire process, from research to final decision.

The same survey found that 43% of respondents had discovered a new brand through AI. It also found that only 20% say a brand stands out because it appears earlier in the answer. What matters, in the survey's words, is how clearly and relevantly a brand is described. These are self-reported answers from people who already use AI tools, so treat them as direction, not a census.

The practical reading: you do not have to win position one in an AI answer. You have to be described clearly enough to make the shortlist, and be easy to verify when the buyer checks.

Where is search heading in 2027?

Nobody can forecast this, so this section sticks to directions the vendors have announced or the data already shows. Three are worth planning around.

  • Agents will research and book for people.
  • There will be several AI tools, not one.
  • Measurement moves from rank to appearance.

Agents will research and book for people

Google has said which way it is going. In its May 2026 announcement, Google wrote: "We're entering the era of Search agents, where you can easily create, customize and manage multiple AI agents for your many tasks, right in Search."

For shopping, OpenAI's help page on shopping in ChatGPT says merchants are ranked based on factors like availability, price, quality, and whether they are the maker or primary seller of that item. It also says product results are not ads. For merchants on Shopify, the same page says product data is already integrated into ChatGPT through Shopify Catalog, and that no additional work is required from individual merchants.

The honest limit is how many shoppers will hand over the purchase. In the Semrush survey above, only 20% use AI from research to final decision today. If you sell on Shopify, our agentic commerce checklist covers the product data, stock, delivery and return fields an agent reads.

There will be several AI tools, not one

One tool does not own this market, and its share is moving. Similarweb's 2026 generative AI report, July 2026, says ChatGPT's share of generative AI website visits fell from about 76% in June 2025 to roughly 53% by May 2026. Gemini rose from under 9% to around 27 to 28% over the same period. These are worldwide web visits, not users.

The tools also treat brands differently. The Semrush and Kevin Indig study found Gemini names brands far more often than it links them, and ChatGPT links far more often than it names them. A result in one tool does not predict the result in another, so a brand needs to look at each one.

Share of generative AI website visits, worldwide

ChatGPT's share fell by more than 20 percentage points in twelve months while Gemini's rose from under 9% to around 27 to 28%.

  • ChatGPT, June 2025about 76%
  • ChatGPT, May 2026about 53%
  • Gemini, June 2025under 9%
  • Gemini, May 202627 to 28%

Source: Similarweb, July 2026. The Gemini June 2025 bar is drawn at 9% and printed as Similarweb states it, under 9%. The Gemini May 2026 bar is drawn at the middle of the stated 27 to 28% range.

Measurement moves from rank to appearance

A ranking tells you where a page sat. An AI answer is different each time it is asked. SparkToro with Gumshoe, January 2026, had 600 volunteers run 12 different prompts through each of 3 tools a combined 2,961 times. They concluded there is a less than 1 in 100 chance that ChatGPT or Google's AI, if asked 100 times, will give the same list of brands in any two responses. The page discloses that a co-researcher works at an AI tracking startup.

That makes one check close to meaningless and a repeated record useful. Google's Search Console help page says that as of August 31, 2026, Google rolled out its generative AI insights to all websites worldwide. It defines impressions as how many times links to your site were shown to a user in a generative AI feature on Google Search.

That report covers Google's features only. For the method, see how to measure AI search visibility.

How should a brand prepare?

A brand prepares by doing five things in order, starting with the cheapest. None of them depends on guessing which AI tool wins.

  1. Keep the basics working.
  2. Write down the questions buyers ask.
  3. Publish facts nobody else has.
  4. Be mentioned by sources other people trust.
  5. Measure appearances as well as rank.

1. Keep the basics working

Keep every page you want used crawlable, indexed, fast to load and accurate. Google's own guidance is that eligibility for its AI features starts with being indexed and snippet-eligible, and that SEO best practices continue to be relevant. Check that prices, stock and delivery times on the page match what a buyer will find at checkout.

Each AI tool has its own rules for crawlers and sources. Our guide on how to get cited in ChatGPT and AI search goes through them one tool at a time.

2. Write down the questions buyers ask

List the real questions customers ask before they buy, in their own words. Support tickets, product reviews and sales calls are the best sources. These become the question set you test, and the headings of the pages you improve.

The Semrush survey found that clarity matters more than position, so the wording buyers use is the wording your pages should answer. Our post on why customers buy shows how to pull that language out of reviews and tickets.

3. Publish facts nobody else has

Add the information only you can add: your own product tests, measured sizes, real delivery times, how returns work in practice, answers from your support team. Google says unique, compelling and useful content will likely matter more than any other suggestion in its guide, and a fact nobody else publishes is the easiest way to be unique.

4. Be mentioned by sources other people trust

AI answers cite pages across the web, not only brand sites. Pew Research Center, July 2025, found that the most frequently cited sources in both Google AI summaries and standard search results are Wikipedia, YouTube and Reddit. Being discussed where buyers read gives AI more to find, though no study we read shows that a mention causes an appearance. The evidence on which sources matter most is in our guide on getting cited in ChatGPT and AI search.

One caution from Google: its guide says that seeking inauthentic "mentions" across the web is not as helpful as it might seem. Earn mentions with something worth mentioning.

5. Measure appearances as well as rank

Run your buyer questions across ChatGPT, Google AI Overviews, Google AI Mode, Gemini and Perplexity, record whether each one names you and links you, and repeat it, because answers change. Our guide on how to measure AI search visibility sets out five measures and a monthly routine you can run yourself.

If you would rather not run it by hand, we do this for ecommerce brands. AI Search runs the buyer questions across those five tools and reports every month, and the first check is free. You can also see how we work together. We cannot promise a brand will be named, because answers vary run to run.

Frequently asked questions

Is SEO still relevant with AI search?
Yes, according to Google. Its generative AI guide says the best practices for SEO continue to be relevant because its generative AI features are rooted in its core Search ranking and quality systems. A page must also be indexed and eligible to show with a snippet to appear as a supporting link in AI Overviews or AI Mode, with no additional technical requirements. That covers Google only, and the other AI tools work differently.
How do AI answers change clicks to websites?
They reduce them in the studies we found. In Pew Research Center, July 2025, users clicked a result link in 8% of visits with an AI summary and 15% without one. A 2026 field experiment on SSRN found AI Overviews cut outbound organic clicks by 39.8%, and an Ahrefs study, February 2026 found a 58% lower average desktop click-through rate for the top-ranking page, as a correlation. Google says the clicks that remain are higher quality.
What is Google AI Mode?
AI Mode is a mode of Google Search that answers questions with an AI-generated response. Google says in its May 2026 Search announcement that it passed one billion monthly users, and said on its Q3 2026 earnings call that its queries are on average three times longer than traditional searches.
Will shoppers buy inside AI tools?
Some may, but the evidence so far says most still use several channels. In a Semrush survey, December 2025, of 1,030 US shoppers who had tried AI tools, only 20% said they use AI from research to final decision. OpenAI says ChatGPT ranks merchants on availability, price, quality and whether they are the maker or primary seller of the item.
Do AI answers change from one check to the next?
Yes. In a SparkToro with Gumshoe study, January 2026, covering 2,961 runs, there was a less than 1 in 100 chance that ChatGPT or Google's AI would give the same list of brands in any two responses. One check proves little, so repeat the same questions and look at how often a brand appears.

Search is still how buyers find brands, but AI answers can decide which names reach the shortlist for some buyers, and no one can promise yours will be named. To see where the AI tools name your brand today, start with the free AI visibility check on our AI Search page.

How well does your store run?

Score six areas of your store, from stock to support, and see the three to fix first.

Take the Free Assessment About 3 minutes. Results instantly.