What 1 million keywords reveal about AI’s impact on search

Search demand is shifting, not shrinking. Our evaluation discovered that 29% of high-volume search demand is in decline, whereas almost the identical quantity is rising elsewhere. Total demand stays basically flat as a result of search habits is being redistributed moderately than lowered.

Focus your search engine optimisation technique on the place demand is rising. Evaluate which key phrases are dropping quantity, determine the queries gaining momentum, and construct the model authority that earns visibility in each search engines like google and yahoo and AI-generated solutions.

This examine examines the place search demand is shifting, which industries are seeing the largest modifications, and what these patterns imply for search engine optimisation.

How we studied AI’s influence on search

In 2024, Gartner predicted that traditional search engine volume would fall 25% by 2026 as shoppers shifted to AI chatbots and digital brokers. Fractl and Search Engine Land got down to check that prediction. (Disclosure: I’m the co-founder of Fractl.)

We analyzed Semrush knowledge for 1,010,848 high-volume key phrases, every with 10,000 or extra month-to-month searches, throughout 379 manufacturers in eight verticals. We additionally surveyed 1,004 U.S. shoppers to grasp how AI is altering search habits.

The evaluation measured which key phrases gained and misplaced search quantity over the previous yr, how these modifications assorted by trade, and the way shopper search habits are evolving alongside AI adoption.

The 29% search decline is actual, but it surely varies so much by vertical

Throughout greater than one million high-volume key phrases, 29% of search quantity is in measurable decline. That’s 4 share factors above Gartner’s forecast. In a dataset representing 35.4 billion month-to-month searches, a 4-point distinction interprets to a considerable quantity of searches.

By trade, FinTech noticed the most important decline at -37.7% and Way of life the smallest at -15.2%. Solely three of the eight verticals (Insurance coverage, SaaS, and Way of life) got here in beneath Gartner’s 25% threshold. FinTech, HealthTech, and Wellness all landed effectively above it.

The sample tracks with how information-heavy a class is. The place a chatbot can present an entire reply, reminiscent of a drug interplay abstract, a proof of deductibles, or a fast fund overview, search quantity declines. Classes that require folks to match costs, full a purchase order, or navigate to a selected website retain extra search demand.

Verticals the place folks have to transact (SaaS, Way of life, Insurance coverage, and Journey) are rising or stay near flat. Verticals the place folks primarily search data (HealthTech, FinTech, and Wellness) are seeing the most important search declines.

Benchmark these findings in opposition to your individual vertical earlier than reacting to AI-driven search declines. HealthTech and FinTech groups ought to plan for publicity effectively above the general 29% decline. SaaS and Way of life groups have extra motive to query broad claims that search demand is collapsing.

Search demand is redistributing

The highest-line decline quantity will get the headlines. The offset issues as a result of it exhibits that demand didn’t vanish. It moved to a distinct set of phrases, and people are those price rating for.

Sure, 40.7% of the high-volume key phrases we tracked are in measurable decline, which means greater than a 15% quantity loss over the previous yr. Among the many affected key phrases, the common decline is -41%, and 112,378 of them have misplaced greater than 40% of their quantity. For manufacturers rating on these phrases, the hit isn’t mild.

However 20.1% of key phrases are rising by that very same threshold. Add up the amount on either side of the ledger, and the 2 totals nearly precisely cancel out.

The 285,489 declining key phrases add as much as roughly 10.29 billion in month-to-month quantity. The 140,835 rising key phrases add as much as roughly 10.31 billion. Web change throughout your complete dataset: +16.8 million searches a month.

Fewer key phrases are rising than declining, however the rising ones carry extra quantity every, which is how the totals find yourself even. Demand didn’t shrink a lot as relocate.

The vertical-level growth-to-decline ratios present the place the brand new demand is touchdown:

• Way of life leads at 2.6x (40% rising vs. 15% declining).
• SaaS is correct behind at 2.5x (48% rising vs. 19% declining).
• HealthTech sits on the different finish with an inverted ratio of 0.4x, which is why it’s probably the most disrupted vertical within the set.

Pull your tracked key phrases, filter by year-over-year quantity change, and see which aspect of the ledger your portfolio sits on. That’s the audit to run first.

Non-branded queries are probably the most weak

AI chatbots substitute non-branded queries simply. When a search time period doesn’t embody a model identify, there’s no specific website the person has to achieve and no particular supply the reply has to return from, so the entire trade can keep contained in the chat window.

Throughout the dataset, 90% of all tracked search quantity is non-branded. HealthTech (99.6%) and Wellness (98.5%) are probably the most uncovered. Insurance coverage (73.8%) and SaaS (82.0%) are much less uncovered, and each are rising general. SaaS quantity is up 48% previously yr, and Way of life is up 40%.

If you wish to know which content material is most in danger, the keyword-pattern knowledge is the clearest sign within the examine.

The rationale SaaS and Way of life may be each AI-targeted and rising comes all the way down to what occurs after AI solutions.

When AI recommends a challenge administration device or a sofa, loads of folks nonetheless seek for the precise model or retailer earlier than they purchase. The AI reply creates a downstream search. In HealthTech and FinTech, it doesn’t. A drug-interaction query or a “what’s a deductible” question will get answered in full contained in the chat window, and there’s no subsequent step that sends you to Google.

In case your class produces full AI solutions with no pure subsequent click on, you want an AI visibility technique, not simply an search engine optimisation one. Exhibiting up within the reply is the entire recreation now.

70% of shoppers use AI extra, however simply 17% use search much less

The key phrase knowledge tells you what’s taking place within the index. The survey tells you what’s taking place within the heads of the folks doing the looking.

Search habits has unfold throughout extra platforms. Loads of persons are folding AI into their habits with out giving up Google.

Social platforms are working as search engines like google and yahoo in a approach they weren’t a number of years in the past. YouTube (68%) and Reddit (57%) lead by a large margin, with Instagram (42%), Fb (40%), and TikTok (33%) rounding out the highest 5.

YouTube and Reddit are price prioritizing when you haven’t already. Each rank forward of TikTok, Instagram, and Fb as search locations, and each index in Google, so visibility there compounds throughout platforms.

What has truly moved from Google to AI

Greater than a 3rd of respondents (35%) say they haven’t changed conventional search with AI for something but. Amongst those that have, how-to guides and tutorials took the largest hit.

For buy analysis, 47% of shoppers begin with a conventional search engine, tied with on-line retailers at 47%. Solely 13% begin with an AI chatbot, and consumers examine a median of three on-line sources earlier than they purchase.

Right here’s the quantity price bringing to your subsequent technique assembly: Almost one in 5 shoppers (18%) have purchased one thing based mostly on an AI advice with out checking it in opposition to a separate search.

That’s a brand new form of purchaser journey, one the place the model by no means will get a search-driven touchpoint. To be within the working, it’s a must to be one of many names the chatbot returns.

Gen Z and millennials are 2.5x extra seemingly than child boomers to purchase based mostly on an unverified AI advice (20% vs. 7%). Throughout all shoppers, 59% say they’re more likely to go to a model’s web site after an AI chatbot mentions or recommends it.

That’s the brand new conversion funnel. Model mentions in an AI reply are the brand new rankings. Visits to a model’s web site are the brand new click-throughs.

On belief:

• 33% of shoppers belief AI and conventional search equally.
• 46% belief search extra.
• 20% belief AI extra.

Greater than half (56%) are at the least considerably skeptical of AI product suggestions. Persons are blissful to let AI filter and shortlist, however most nonetheless confirm earlier than they purchase.

The 5-year outlook: Google’s not going wherever, however the minority that’s leaving issues

Requested whether or not Google will nonetheless be their main search device in 5 years, 52% of shoppers say sure (17% undoubtedly, 35% in all probability). One other 27% aren’t positive, and 20% say in all probability or undoubtedly not.

The highest causes folks desire AI over conventional search are higher summaries throughout sources (21%), quicker and extra direct solutions (20%), and the power to ask conversational follow-up questions (19%). Extra personalised outcomes and never having to click on via web sites trailed far behind, at 6% and 4%.

Requested what would convey them again to conventional search, the highest reply was AI giving unreliable solutions (35%), a lot of this cut up is dependent upon whether or not AI maintains folks’s belief because it scales. Extra correct search outcomes got here subsequent (29%), adopted by a desire for a number of supply hyperlinks (22%) and privateness considerations (20%).

The 20% who anticipate to go away Google aren’t a majority, however they’re not a rounding error both. You don’t should rebuild your entire technique round them as we speak, however you do want to indicate up the place they’re already going.

What this implies to your content material and search engine optimisation technique

Gartner’s 25% prediction was the proper of directional warning. The actual shift is steeper, however calling it a “decline” misses the larger story. Whole search quantity is mainly flat. What’s modified is which searches carry the amount.

AI visibility is a distribution channel, not a risk to dodge. With 59% of shoppers saying they’ll go to a model’s web site after an AI point out, GEO is now a significant a part of model discovery. Earned media, credible third-party protection, and entity indicators assist manufacturers land in chatbot solutions, which is why digital PR and GEO more and more overlap.

Search is shifting, not disappearing.

The manufacturers that lose are nonetheless optimizing for queries AI now solutions higher. The manufacturers that win are constructing the form of authority that makes them the reply, whether or not it comes from Google or a chatbot.

Methodology

This examine mixed two knowledge sources to check Gartner’s 2024 prediction that conventional search engine quantity would fall 25% by 2026.

Fractl and Search Engine Land analyzed Semrush search quantity knowledge for 1,010,848 high-volume key phrases with 10,000 or extra month-to-month searches every, protecting 379 manufacturers throughout eight verticals: FinTech, HealthTech, Wellness, Journey, Schooling, Insurance coverage, SaaS, and Way of life. The dataset represented 35.4 billion in mixture month-to-month search quantity. Key phrase-level year-over-year quantity change was measured as of April 2026 and labeled as declining (greater than 15% loss), steady (inside 15%), or rising (greater than 15% acquire). Question sample groupings (“What’s X,” “Greatest X for Y,” “X vs. Y,” and “Find out how to X”) had been utilized on the key phrase degree.

Fractl and Search Engine Land additionally surveyed 1,004 U.S. shoppers about their search habits, AI device adoption, and buy analysis habits. The pattern was 52% ladies, 46% males, and 1% nonbinary, with 49% millennials, 26% Gen X, 16% Gen Z, and 9% boomers. The median respondent age was 41, with a spread of 18 to 82.