A widely cited Gartner forecast has already been surpassed. Search volume across more than a million high-volume keywords declined 29 percent over the past year, four points beyond the 25 percent drop Gartner projected for 2026. But the overall figure, published July 2, 2026, obscures a study that found something stranger underneath it: aggregate search demand hasn't shrunk. It has moved.

Growth marketing agency Fractl, working with Search Engine Land, released the findings in a report titled "Search Isn't Dying, It's Redistributing." The study pulled Semrush data on 1,010,848 high-volume keywords, each drawing at least 10,000 monthly searches, spanning 379 brands across eight verticals. It paired that keyword analysis with a survey of 1,004 U.S. consumers conducted to understand how AI tools are reshaping search behavior.

A forecast tested, and exceeded, in one direction

In February 2024, Gartner analyst Alan Antin forecast that traditional search engine volume would fall 25 percent by 2026 as consumers shifted toward AI chatbots and virtual agents. The prediction spread quickly through trade press and, according to the report, likely crossed many a CMO's desk in internal AI-search strategy memos over the following eighteen months.

Fractl set out to test that number against real keyword data rather than survey sentiment. The team found the prediction directionally correct but, on the surface, more severe: 29 percent of tracked search volume showed measurable decline as of April 2026.

Yet the more consequential finding sits one layer beneath that headline. Two distinct groups of keywords moved in opposite directions, and when their volumes are added together, the totals nearly cancel out. Some 285,489 declining keywords represented approximately 10.29 billion in monthly search volume. Meanwhile, 140,835 growing keywords represented roughly 10.31 billion. The net change across the full dataset came to positive 16.8 million searches per month, a figure the report calls "essentially flat" relative to the scale of the data. Fewer keywords are gaining volume than losing it, but the winning keywords each carry more traffic on average, which is how the ledger balances.

That framing changes the strategic question facing marketing teams. Rather than asking how to slow a decline, the report argues that teams should be asking whether their content targets the keywords still gaining ground, or the ones already losing it.

Eight verticals, one wide gap

The 29 percent aggregate figure masks substantial variation by industry. Three of the eight verticals Fractl analyzed, Insurance, SaaS, and Lifestyle, came in below Gartner's 25 percent threshold, while FinTech, HealthTech, and Wellness landed well above it.

FinTech recorded the steepest decline in the dataset at negative 37.7 percent; Lifestyle recorded the smallest at negative 15.2 percent. That gap, more than 22 percentage points, is wider than the difference between Gartner's prediction and the study's overall finding.

The pattern tracks how information-heavy a category is. Where a chatbot can deliver a complete, self-contained answer, such as summarizing a drug interaction or explaining an insurance deductible, search volume for that query type declines. Where the task instead requires comparing prices, navigating to a specific site, or completing a transaction, search volume holds or grows.

Verticals built around transactions, including SaaS, Lifestyle, Insurance, and Travel, are growing or near flat. Verticals built around explaining concepts, including HealthTech, FinTech, and Wellness, show the largest declines. The vertical growth-to-decline ratios make the divide explicit. Lifestyle leads at 2.6x, with 40 percent of tracked keywords growing against 15 percent declining. SaaS follows at 2.5x, with 48 percent growing against 19 percent declining. HealthTech sits at the opposite extreme, inverted at 0.4x, with 37 percent declining against just 14 percent growing, making it the most disrupted vertical measured.

Branded queries hold, non-branded queries don't

A second cut of the data explains why the vertical split looks the way it does. Across the full dataset, 90 percent of tracked search volume is non-branded, meaning queries that don't reference a company name, such as "what is a 401k" or "best running shoes for flat feet." These carry no requirement that the answer come from a particular destination, so the exchange can stay entirely inside a chat window.

HealthTech and Wellness carry the highest non-branded exposure, at 99.6 percent and 98.5 percent respectively. Insurance and SaaS carry the lowest, at 73.8 percent and 82.0 percent, and both verticals are growing overall. Insurance's resilience traces partly to a 26 percent branded share: a search for a specific insurer's claims process requires a destination a generic chatbot answer cannot replace.

The implication is direct. The 90 percent of volume classified as non-branded overlaps heavily with the informational, top-of-funnel content that SEO and content teams have built strategies around for roughly a decade, the category the study identifies as most exposed to AI displacement. Branded query share, by contrast, is the asset that holds its value.

The middle of the funnel, not the top, is collapsing fastest

Conventional wisdom about AI search displacement assumes a top-down pattern: chatbots absorb awareness-stage queries first, then gradually work toward transactional intent. Fractl's keyword data contradicts that sequence.

Top-of-funnel informational queries showed the smallest decline, at 19 percent. Bottom-of-funnel transactional queries came in higher, at 34 percent. The steepest decline appeared in the middle: comparison queries, the "X vs. Y" and "best X for Y" searches tied to the consideration stage, fell 36 percent.

Comparison prompts, the report argues, are exactly what large language models answer well. A prompt such as "best CRM for small business" or "Toyota Camry vs Honda Accord" returns a synthesized, side-by-side answer inside the chat interface, without the user opening several review sites to compare manually.

The pattern varies by vertical. FinTech shows the most extreme funnel compression, with comparison queries losing volume fastest of any type measured. Education shows a reversal, with bottom-of-funnel queries declining faster than top-of-funnel ones, which the report attributes to enrollment and scheduling increasingly happening through AI conversation. HealthTech shows bottom-of-funnel queries declining sharply while top-of-funnel queries hold stable, tied to health searches remaining anchored to trusted destinations such as WebMD or the Mayo Clinic in a way comparison searches are not.

Four query patterns driving the shift

The study identifies four specific query structures accounting for most of the observed displacement, each concentrated in a different vertical:

  • Definitional queries ("what is X"): AI provides a direct conversational answer. Most affected vertical: HealthTech.
  • Listicle queries ("best X for Y"): AI gives a direct recommendation. Most affected vertical: Lifestyle, even though Lifestyle is growing overall.
  • Comparison queries ("X vs. Y"): AI synthesizes trade-offs instantly. Most affected verticals: SaaS and FinTech, though SaaS is also growing overall.
  • Tutorial queries ("how to X"): AI walks users through steps conversationally. Most affected vertical: Education.

Decline rates for these four patterns, by vertical, ranged from negative 29 percent in FinTech to negative 13 percent in SaaS, with Travel at negative 27 percent, Wellness at negative 24 percent, Insurance at negative 21 percent, HealthTech at negative 20 percent, Education at negative 19 percent, and Lifestyle at negative 14 percent.

The counter-narrative, and the reason SaaS and Lifestyle can register heavy AI displacement on these specific query patterns while still growing overall, comes down to what happens after the AI answer arrives. A user who asks a chatbot to recommend project management software still tends to search the recommended brand names directly. A user who asks AI to find a cheap couch still tends to search specific retailers and product names before buying. In HealthTech and FinTech, that downstream search step largely disappears, since a definitional health or finance answer is often complete on its own.

The keyword data shows what is happening inside the search index. Fractl's consumer survey shows what is happening in the habits of the people typing the queries, and the two tell slightly different stories. Seventy percent of the 1,004 respondents said they use AI tools more than a year earlier, yet only 17 percent said they use traditional search less, suggesting the two behaviors mostly coexist rather than one replacing the other.

Generational patterns varied. Gen X reported the largest year-over-year increase in AI usage at 74 percent, with millennials and baby boomers close behind at 72 percent each. Gen Z reported the smallest increase, at 57 percent, which the report attributes to Gen Z already using AI tools at a higher baseline, leaving less room to grow.

Search behavior has also spread across platforms functioning as search engines in practice. YouTube led at 68 percent and Reddit followed at 57 percent, with Instagram at 42 percent, Facebook at 40 percent, and TikTok at 33 percent completing the top five.

Asked which tasks they had shifted from Google to AI tools, respondents ranked how-to guides and tutorials highest at 32 percent, followed by health and medical information at 29 percent, product and shopping research at 25 percent, recipe and food ideas also at 25 percent, financial and legal questions at 24 percent, travel planning and customer service both at 20 percent, entertainment recommendations at 17 percent, news and current events at 16 percent, and local business lookup at just 10 percent. Some 35 percent said they had not replaced search with AI for any task measured.

An emerging zero-search purchase path

Nearly half of consumers, 47 percent, still begin purchase research with a traditional search engine, tied exactly with the 47 percent who begin at an online retailer. Only 13 percent start with an AI chatbot, and the average consumer visits three online sources before buying.

Within that picture sits a smaller, distinct pattern: 18 percent of consumers reported making a purchase based on an AI recommendation without separately verifying it through search. It is a minority behavior in aggregate, but it describes a purchase path in which a brand never receives a search-driven touchpoint at all; to be considered, the brand must be among the names the AI tool surfaces directly.

Gen Z and millennial respondents were 2.5 times more likely than baby boomers to buy on an unverified AI recommendation, at 20 percent versus 7 percent, and Gen Z showed the highest downstream engagement, with 19 percent saying they were "very likely" to visit a brand's website after a mention. Across all age groups, 59 percent said they were at least somewhat likely to visit a brand's website following an AI mention.

Trust remains more cautious than usage suggests. Thirty-three percent of consumers trust AI and traditional search equally, 46 percent still trust search more, and 20 percent trust AI more, while 56 percent described themselves as at least somewhat skeptical of AI product recommendations specifically.

Asked where their purchase journey concludes, respondents pointed most often to online retailers at 46 percent, followed by brand websites at 23 percent, in-store purchases at 17 percent, social commerce at 8 percent, and buying directly through an AI tool at 5 percent, a figure the report frames as the leading edge of the unverified-purchase behavior above.

A five-year outlook that splits roughly in half

Asked whether Google will remain their primary search tool in five years, 52 percent answered affirmatively, split between 17 percent who said "definitely" and 35 percent who said "probably." Another 20 percent said probably or definitely not, and 27 percent were unsure.

Among consumers who prefer AI over traditional search, the leading reasons were better summarization across sources at 21 percent, faster and more direct answers at 20 percent, and the ability to ask conversational follow-ups at 19 percent, while personalized results and avoiding website clicks trailed well behind at 6 percent and 4 percent. More than a quarter, 28 percent, said they do not prefer AI over traditional search at all.

What would bring consumers back to search matters most for how this settles. AI providing unreliable answers ranked first at 35 percent, followed by more accurate search results at 29 percent, a preference for multiple source links at 22 percent, privacy concerns with AI at 20 percent, and AI becoming too expensive at 16 percent. The split, the report notes, will depend heavily on whether AI-generated answers remain reliable as usage scales.

Context from PPC Land's ongoing coverage

The Fractl findings arrive as PPC Land has tracked a wider reckoning with how AI-mediated discovery redistributes attention, traffic, and revenue across search and publishing.

Some of that reporting anticipated the mechanism Fractl's data now quantifies. Google AI Overviews reduce organic CTR 61%, paid traffic 68% covered Seer Interactive research, published November 2025, finding organic click-through rates on informational queries fell from 1.76 percent to 0.61 percent once AI Overviews appeared, while paid click-through rates on the same queries dropped from 19.7 percent to 6.34 percent, consistent with the HealthTech and FinTech exposure Fractl documents.

The publisher side of that dynamic surfaced separately. Small publishers lost 60% of search traffic as AI reshapes the webcovered Chartbeat data, published March 2026, showing outlets with 1,000 to 10,000 daily page views lost 60 percent of search referral traffic over two years, while ChatGPT referrals stayed under 1 percent of total publisher page views. Fractl's finding, that demand redistributes toward non-branded content rather than vanishing, helps explain why that loss concentrated on smaller, informational sites.

The zero-search purchase behavior Fractl documented connects to a broader body of research on AI-influenced commerce. Your analytics are lying: Similarweb traces AI recommendations to real traffic found AI-recommended brands were 2.5 times more likely to receive a site visit within seven days, with 56 percent of that traffic arriving through branded search rather than a direct AI click, a pattern consistent with the "new conversion funnel" Fractl describes.

Brand visibility inside AI answers has become its own measurement discipline over the same period. Semrush: 36 brands win AI visibility everywhere, 1,200 vanish on one covered an Adobe-owned Semrush analysis of 126 million AI search prompts, finding that of more than 1,200 brands tracked across 22 verticals, only 36 achieved consistent top-100 visibility across every platform studied. That gap underscores why Fractl's finding, that 59 percent of consumers visit a brand's site after an AI mention, carries direct commercial weight.

Why this matters for marketers

Fractl's report makes a reallocation argument rather than a defense argument. Net search demand across the tracked dataset moved only 16.8 million searches per month, a rounding error against the 35.4 billion in aggregate monthly volume the underlying keywords represent. Brands losing ground in the study are generally still optimizing content for queries AI tools now answer more completely than a results page does. Brands gaining ground are those already ranking for branded and transactional queries that continue sending users to a specific destination rather than resolving fully inside a chat window.

For marketing teams weighing budget against this data, the vertical figures matter more than the 29 percent headline. A HealthTech or FinTech program built around definitional and comparison queries sits in the most exposed segment of the dataset. A SaaS or Lifestyle program built around the same query types may lose volume on those specific patterns while still growing in aggregate, since AI-driven discovery in those categories routes a meaningful share of users back toward a branded search a step later.

Timeline

  • February 19, 2024: Gartner publishes its forecast that traditional search engine volume will fall 25 percent by 2026 due to AI chatbots and other virtual agents.
  • April 2026: Fractl and Search Engine Land conduct the keyword-level year-over-year volume analysis underlying the study, covering 1,010,848 high-volume keywords across 379 brands in eight verticals.
  • April 2026: Fractl surveys 1,004 U.S. consumers on search habits, AI tool adoption, and purchase research behavior.
  • July 2, 2026: Search Engine Land publishes the full study, authored by Fractl co-founder Kelsey Libert, under the title "What 1 million keywords reveal about AI's impact on search."

Summary

Who: Fractl, a growth marketing agency, working with Search Engine Land and drawing on Semrush search volume data.

What: An analysis of 1,010,848 high-volume keywords across 379 brands in eight verticals found overall search volume declined 29 percent over the past year, four points beyond a 2024 Gartner forecast, though the decline is offset almost entirely by growth in a separate set of keywords, leaving aggregate demand roughly flat. A companion survey of 1,004 U.S. consumers found 70 percent report using AI tools more than a year ago, while only 17 percent report using traditional search less.

When: The keyword and survey data were measured as of April 2026. Search Engine Land published the full study on July 2, 2026.

Where: The keyword analysis covered U.S.-relevant search volume across eight industry verticals. The consumer survey sampled 1,004 respondents across the United States.

Why: The findings challenge a simple narrative of search decline by showing that demand is redistributing rather than shrinking, with sharply different consequences by vertical, funnel stage, and query type. For marketing teams allocating content and search budgets, the report argues the practical question is not whether to defend against AI displacement broadly, but which specific keyword categories, within their own vertical, are gaining volume and which are already lost.