The Macro Shift · contested evidence
Did the Search Apocalypse Arrive? Auditing Gartner's 25%-by-2026 Prediction
In February 2024, Gartner published the single most repeated forecast in this field: that traditional search engine volume would fall 25 percent by 2026 as buyers moved their questions to AI chatbots and virtual agents. That Gartner search prediction quickly became shorthand for a coming search apocalypse. Two years on, it is fair to audit it against the record, and the record is clear. Search volume did not collapse 25 percent, and Google still holds more than 90 percent of the search market. What did change is real but different from what the headline promised: the click eroded, AI answers now sit inside Google's own results, and buyer attention spread across more surfaces rather than draining away from search. Saying this plainly matters, because the conclusion is neither denial nor doom. The forecast was wrong as literally stated, the underlying shift is genuine, and the right response is to measure where a business stands across every surface, not to react to a prediction that did not come true.
The forecast, stated precisely
Auditing a prediction requires quoting it exactly, because most of the confusion in this debate comes from paraphrase. On 19 February 2024, Gartner issued a press release titled "Gartner Predicts Search Engine Volume Will Drop 25% by 2026, Due to AI Chatbots and Other Virtual Agents." The claim was specific in three ways: a magnitude (25 percent), a deadline (2026), and a mechanism (query volume migrating from search engines to AI chatbots and virtual agents).
Specificity is a virtue in a forecast, because it makes the forecast testable. A vague claim that "search is changing" can never be falsified. A claim that a named quantity will move by a named percentage by a named year can be checked. That is the standard we hold it to here, and it is the standard the firm implicitly invited by publishing a number rather than a mood.
What actually happened by 2026
Reporting that checked the prediction against 2026 reality reached a blunt conclusion: traditional search volume has not dropped 25 percent, and Google still holds more than 90 percent of the search market. The migration the forecast described, a bulk transfer of queries away from search engines and toward standalone chatbots, is not the pattern the data shows.
The pattern that did emerge is closer to absorption than exodus. Rather than losing queries to competitors, Google embedded the generative-answer experience inside its own results as AI Overviews. The query still runs on the search engine; it simply returns a synthesized answer above the links. On that reading, the AI-search shift is happening largely within search, not as a replacement of it, which is precisely why a headline about collapsing search volume misses what changed.
So is search dying, and will AI replace Google?
The answer to both questions is no, and the distinction is the difference between a click and a query. Buyer attention is being reallocated across more surfaces, not disappearing from search. The click, the act of leaving the results page for a website, has genuinely eroded. The query, the act of asking a search engine, has not migrated en masse to a Google-killer.
This is why the framing matters. "Is search dying" and "will AI replace Google" are the wrong questions if you take them literally, because they invite a yes-or-no answer to a phenomenon that is really a change of shape. The useful question is narrower and answerable: for a given buyer's question, on the surfaces that question now reaches, is a business present or absent?
The click did erode, and that part is well evidenced
The strongest primary anchor here is the Pew Research Center's July 2025 study, which tracked the actual browsing of 900 consenting US adults across 68,879 Google searches. Pew found that users clicked through to a traditional result in 8 percent of searches that showed an AI summary, versus 15 percent for searches without one, and that only 1 percent of visits went to a link cited inside the AI summary itself. Independent clickstream analysis points the same direction: SparkToro, using Datos and later Similarweb panel data, reported US zero-click searches rising from 58.5 percent in its 2024 study to 68.01 percent across early 2026.
So the erosion of the click is real, measured, and consistent across independent methods. That is the part of the "search is changing" story that survives scrutiny. It is also a very different claim from a 25 percent fall in search volume.
But the query stayed on the search engine
The same SparkToro and Similarweb analysis found that AI Overviews now appear on more than 20 percent of Google searches and, when present, cut click-through rate by close to 60 percent. Read carefully, that is a statement about what happens to clicks on Google, not about queries leaving Google. The engine is answering more and sending fewer clicks, which suppresses referral traffic while leaving Google's position as the place people ask intact. A forecast about volume draining to chatbots and a reality of clicks being absorbed inside Google are not the same event.
Why confident search-volume forecasts keep failing
The Gartner miss is not an isolated embarrassment; it is characteristic of a domain where the measurement itself is unsettled. Current estimates of how large the AI-search shift already is vary by an order of magnitude depending on who is counting and what they count. In a single research pass, ChatGPT has been reported at roughly 17 percent of global digital queries by one source, near 20 percent of global search-related traffic by another and 12 percent in the US, and as commanding 92.4 percent of trackable large-language-model referral traffic by a third measuring an entirely different denominator. A B2B-specific study put ChatGPT's share of measurable AI referrals at 62.6 percent as rival assistants grew.
When credible figures for the same phenomenon range from 12 percent to more than 90 percent, no standardized, transparent measurement of AI-search share exists yet. Even a single product's behavior moves sharply: ChatGPT's own search feature reportedly triggered on 34.5 percent of its queries in February 2026, down from 46 percent in late 2024. A forecast built on top of numbers this unstable inherits their instability. That is the structural reason a precise 25-percent-by-2026 call was unlikely to land, and the reason we treat every single-vendor AI-search statistic as provisional and always name its denominator.
What the forecast got right, and what it got wrong
A fair audit credits what the prediction saw clearly. Gartner correctly identified the direction of travel: generative AI would reshape how people get answers, and organizations would have to rethink a visibility strategy built purely on ranked links. That directional read has aged well.
What it got wrong was the shape and the arithmetic. It framed the change as substitution, queries leaving search for chatbots, when the dominant mechanism has been absorption, AI answers appearing inside search. And it attached a false precision, a specific percentage by a specific year, to a transition whose own base measurements are not yet reconciled. The lesson is not that forecasting is worthless. It is that in this field the direction is knowable and the timelines and totals are where confident predictions keep going wrong, which is exactly why measurement should replace assertion.
The conclusion, and what it changes
Put the record together and a defensible position emerges. The apocalypse framing was wrong: search did not collapse, and Google still holds the large majority of the market. The complacent framing is also wrong: the click really has eroded, AI answers really do sit above the links on a large and growing share of searches, and reputation and local presence still decide who gets chosen. Attention has fragmented across surfaces rather than vanished from one.
That conclusion has a direct operational consequence. If visibility is spread across several structurally different surfaces at once, classic search, the local map pack, AI answers, and reputation, then a single ranking number cannot describe where a business stands, and a panic-driven pivot to any one surface is a bet, not a plan. What the evidence supports instead is a measured read across all of them, dated and disclosed, expressed as a multi-surface figure rather than a headline. Measuring where you stand is the clear alternative to reacting to a forecast that did not come true.
The evidence
Key findings, with their sources
-
Gartner forecast that traditional search engine volume would drop 25 percent by 2026 due to AI chatbots and virtual agents. As of 2026 that decline has not materialized as stated.
contested Gartner, Inc., "Gartner Predicts Search Engine Volume Will Drop 25% by 2026, Due to AI Chatbots and Other Virtual Agents", press release, 19 Feb 2024.
-
Traditional search volume has not fallen 25 percent, and Google still holds more than 90 percent of the search market; the change has been AI answers embedded inside Google rather than volume collapsing to competitors.
contested Future Factors, "Gartner Said Search Would Drop 25% in 2026. It Didn't." (futurefactors.ai), 2026 reality-check.
-
Users clicked a traditional result in 8 percent of searches with an AI summary present, versus 15 percent without, and clicked a link inside the summary only about 1 percent of the time.
established Pew Research Center, "Do people click on links in Google AI summaries?", 22 Jul 2025 (browsing panel, 900 US adults, 68,879 searches).
-
US zero-click Google searches rose from 58.5 percent in 2024 to 68.01 percent across early 2026.
established SparkToro, "2024 Zero-Click Search Study" (Datos) and "In 2026, Less than One Third of Google Searches Still Send a Click" (Similarweb).
-
AI Overviews appear on more than 20 percent of Google searches and, when present, cut click-through rate by close to 60 percent.
established SparkToro / Similarweb analysis, 2026, covered by Search Engine Land.
-
Credible estimates of ChatGPT's share of search-related activity range from about 12 percent to more than 90 percent depending on the denominator counted; no standardized measurement yet exists.
contested Semrush, Similarweb and industry trade sources, 2026 (denominators differ; cited to document the spread, not a settled figure).
Calibration
What is proven, what is promising, what is unproven
| Evidence tier | Tactics | What the evidence says |
|---|---|---|
| Established | The click has eroded (zero-click rise, lower CTR under AI summaries) | Pew Research Center 2025 (68,879 searches); SparkToro/Datos 2024 and SparkToro/Similarweb 2026 |
| Established | AI answers now appear on a large, growing share of searches | SparkToro/Similarweb 2026 (AI Overviews on 20 percent-plus of searches) |
| Contested | The forecast 25 percent collapse in search volume by 2026 | Gartner 2024 press release vs 2026 reality-check; Google still 90 percent-plus share |
| Contested | How large the AI-search shift already is | Vendor figures spanning roughly 12 percent to 90 percent-plus on different denominators, 2026 |
Reference
Glossary
- Search volume
- The total count of queries run on a search engine. Gartner's 2024 forecast was about this quantity falling 25 percent, which is distinct from clicks or referral traffic falling.
- Zero-click search
- A search that ends without a click to any external website, because the answer is satisfied on the results page. Its rise is well evidenced and is separate from any change in search volume.
- AI Overview
- Google's synthesized answer shown above the list of links. Its spread means AI answers largely appear inside Google rather than pulling queries away to standalone chatbots.
- How often a business is named inside the synthesized answers engines return for a fixed set of buyer questions, sampled repeatedly and reported as a rate rather than a single ranking.
- Falsified as stated
- A forecast whose specific, testable claim did not occur as worded, even if the broader direction it pointed to is real. The Gartner prediction is falsified as literally stated.
Straight answers
Frequently asked questions
Did Gartner's prediction that search volume would drop 25 percent by 2026 come true?
No. Gartner's February 2024 press release forecast a 25 percent decline in traditional search volume by 2026 due to AI chatbots. As of 2026 that collapse has not materialized as stated, and Google still holds more than 90 percent of the search market. The direction the forecast pointed to, generative AI reshaping how answers are delivered, is real; the specific magnitude and timeline were not.
Is search dying?
No. The evidence supports evolution, not extinction. The click has genuinely eroded, with independent studies showing zero-click searches above 60 percent and lower click-through when an AI summary appears, but the query itself still overwhelmingly runs on search. Attention is being reallocated across more surfaces, not vanishing from search.
Will AI replace Google?
Not on the current record. The dominant pattern has been Google absorbing the generative-answer experience as AI Overviews inside its own results, not queries migrating to a standalone competitor. AI answers are changing what a search returns more than they are replacing where people search.
What is Google's market share in 2026?
Reporting that checked the 2024 forecast against 2026 reality found Google still holds more than 90 percent of the search market. That figure is why the 25 percent volume-collapse forecast is treated as falsified as literally stated.
If the forecast was wrong, is AI search overhyped?
Not overhyped, mismeasured. The real, well-evidenced changes (eroding clicks, AI answers on a growing share of searches) sit alongside genuinely contested figures for how large AI search already is, which currently range from about 12 percent to more than 90 percent depending on the denominator. The clear response is to measure standing directly rather than trust any single headline number.
Provenance
Sources
- Gartner, Inc., "Gartner Predicts Search Engine Volume Will Drop 25% by 2026, Due to AI Chatbots and Other Virtual Agents", press release, 19 Feb 2024 (contested, used as a forecast-accuracy check)
- Future Factors, "Gartner Said Search Would Drop 25% in 2026. It Didn't." futurefactors.ai, 2026 (contested reality-check)
- Pew Research Center, "Do people click on links in Google AI summaries?", 22 Jul 2025 (established, primary panel data, 68,879 searches)
- SparkToro, "2024 Zero-Click Search Study" with Datos, sparktoro.com (established)
- SparkToro, "In 2026, Less than One Third of Google Searches Still Send a Click" with Similarweb, sparktoro.com; covered by Search Engine Land (established)
- Semrush, "ChatGPT traffic analysis: Insights from 17 months of clickstream data", 2026 (contested, denominator-dependent)semrush.com
- Similarweb, "Gen AI Stats 2026", and industry trade aggregators, 2026 (contested, cited to document the measurement spread)
Every figure above is attributed to a real, dated source and tagged with its evidence tier. Where a claim could not be verified to a primary source, it is not stated as fact.