The Visibility Corpus

Reading the Corpus

Search reach just hit its lowest point on record, and half of Americans now ask a chatbot instead, on a surface nobody is independently measuring yet.

Original research by Chandranshu Kumar, Founder, Raveneye Global. Published 2026-07-28. · 13 min read

Part of The Attention Landscape in the Insights library.

Abstract

The surface where people look for you is not one surface anymore. Search engine reach among online adults slipped to 80.3% in October 2025, the lowest point in recent tracking, according to DataReportal's Digital 2026 Global Overview Report, while Pew Research Center found that 49% of US adults have now used an AI chatbot and 42% use one specifically to search for information. The rest of the evidence agrees on the direction, even where the surfaces (social, audio, short video, AI answers) don't yet share a common ruler, and the newest of those surfaces has no independent, audited measurement of who gets cited inside it at all.

80.3% (-210 bps YoY), lowest on record Monthly search engine use among online adults DataReportal, Digital 2026 Global Overview Report, Oct 2025
49% have used a chatbot; 42% use it to search for information AI chatbot adoption among US adults Pew Research Center, Americans and AI 2026
49% (up from 39% in 2019) Share of US adults who find news incidentally rather than seeking it Pew Research Center, Apr 2026
6.75 platforms/month; 2.5+ hrs/day on social and video combined Average platform spread and time spent DataReportal, Digital 2026 Global Overview Report, Oct 2025
up to 40% visibility lift Effect of Generative Engine Optimization techniques on citation inside AI answers Aggarwal et al., Princeton / Georgia Tech / Allen Institute for AI (arXiv:2311.09735)
How the market is evolving

The clearest signal in the data is a shift in kind, not just degree. Monthly search engine use among online adults fell to 80.3% in October 2025, a drop of roughly 210 basis points from a year earlier and the lowest level DataReportal's Digital 2026 Global Overview Report has recorded. That single number understates what's happening underneath it: the same report found the average person now spends more than two and a half hours a day across social and video platforms combined, spread across 6.75 different platforms a month, and Pew Research Center clocked 49% of US adults as having used an AI chatbot at all, up from roughly a third the year before. Outside the US the pattern moves faster in places. The Reuters Institute's Digital News Report 2025 found TikTok is already a primary news source for 49% of people in Thailand and 48% in Malaysia, while Brazil's print news reach has fallen from 50% in 2013 to just 10% today and 9% of Brazilians already say they get news from an AI chatbot. None of these numbers share a stopwatch or a sample frame (Pew works from a US probability panel, the Reuters Institute from a 47-country online panel, DataReportal from aggregated platform and traffic data), so read them as a consistent direction, not a single equation: attention that used to concentrate on one search box is now distributed across a dozen surfaces, and the newest of those surfaces is growing from a standing start.

What it does to buyers

What this does to how people actually find and trust a business is more interesting than the raw reach numbers. Pew found that 42% of US adults who use chatbots do so specifically to search for information, and 60% now say they read the summary a chatbot or search engine writes at the top of a results page, meaning a meaningful share of 'searching' no longer produces a list of links at all, it produces one synthesized answer. At the same time, Pew's separate reading-habits research found that 49% of US adults now say they mostly get news by happening to come across it rather than going looking for it, up from 39% when Pew first asked in 2019, and that incidental discovery is far more common among 18 to 29 year olds (52%) than adults 65 and older (28%). Put together, this describes a buyer who is less often typing a query into a box and more often being handed an answer or a headline inside a feed they weren't actively searching. Trust forms earlier and across more small moments than it used to. That is not a new idea (P&G's A.G. Lafley described the First and Second Moments of Truth back in 2005, and Google extended it with the Zero Moment of Truth in 2011, citing 88% of US shoppers researching online before buying), but the number of moments has multiplied and several of them now happen inside a chat window or a synthesized summary rather than on a page you control.

What it means for the attention terrain

For a business owner, the takeaway is that there is no longer one place to win and no single scoreboard that already measures the newest place. Search share trackers still work. Pew and GWI-fed panels still track social and platform reach reliably (YouTube reaches 84% of US adults, Facebook 71%, Instagram 50%, TikTok 32%, per Pew's Social Media Fact Sheet). Edison Research still tracks audio with a probability panel (81% of Americans 12 and older, an estimated 233 million people, listened to online audio in the past month; podcasts reach 58%, or 167 million). But there is no equivalent standing panel or share tracker for how often you get mentioned inside a ChatGPT answer, an AI Overview, or a Perplexity response. DataReportal's own estimate of generative AI platforms crossing a billion combined monthly users comes with the researchers' own caveat that 'an exact figure remains elusive,' stitched together from several partial counts rather than one audited number. That gap, not any single statistic above, is the reason a market's attention terrain has to be mapped surface by surface and re-measured on a real cadence rather than assumed from last year's channel mix. It's also exactly the gap a Machine-Readiness Score is built to fill for your own market.

The data, in one read

ChatGPT weekly active users, 2023-2026
50
Jan 2023
300
Dec 2024
800
Oct 2025
900
Feb 2026
emergingDemandSage's compilation of OpenAI-disclosed weekly-active-user figures shows an eighteenfold rise in three years. The newest surface in this study's terrain map is growing from a standing start, though the billion-user milestone often projected from this curve traces to a third-party analytics estimate, not an OpenAI disclosure. Source: DemandSage, compiling OpenAI-disclosed WAU figures.

The search box is no longer the whole story

For twenty years, a business's discoverability question had one answer: rank in search. That's no longer the whole answer. DataReportal's Digital 2026 Global Overview Report puts monthly search engine use among online adults at 80.3% as of October 2025, down roughly 210 basis points from the year before, the lowest figure the report has tracked. Eighty percent is still a large majority. This isn't a story about search disappearing. It's a story about search no longer being the default first stop it once was, with the difference going somewhere else.

Where does the missing share go? The same report finds the average online adult spending more than two and a half hours a day, 18 hours and 36 minutes a week, across social networks and video platforms combined, and doing it across 6.75 different platforms a month rather than settling on one. Attention isn't consolidating anywhere. It's spreading thinner across more surfaces at once, which is a harder thing to measure and a harder thing to win than a single search ranking ever was.

Eighty percent is still a large majority. It's the direction, not the number, that should change how you think about being found.

A terrain map of where attention actually sits

Ask ten business owners where their customers spend time online and you'll get ten different guesses, most of them a few years out of date. Pew Research Center's Social Media Fact Sheet, drawn from a US probability sample of 5,022 adults surveyed between February and June 2025, gives a more current picture: YouTube reaches 84% of US adults, Facebook 71%, Instagram 50%, Pinterest 37%, TikTok 32%, Snapchat 26%, LinkedIn 25%, WhatsApp 21%, and X just 8%. That's a wide spread, and it means the old shorthand of 'social media' as one channel stopped being useful a while ago; each of those platforms behaves like its own surface with its own audience and its own rules.

Audio tells a similar story with its own separate measurement tradition. Edison Research's Infinite Dial 2026, fielded from a probability-sampled panel in January 2026, found that 81% of Americans age 12 and up, an estimated 233 million people, listened to online audio in the past month, with podcasts alone reaching 58% (167 million). That's a large, stable surface that has almost nothing to do with search rankings or social feeds, and most attention-mapping exercises skip it entirely.

None of these figures share a sampling method or a survey window; Pew, DataReportal, and Edison each measure their own surface on their own cadence. Treat this as a directional terrain map, not a single spreadsheet you can add up to 100%.

The newest surface is growing from a standing start

The most consequential shift in this evidence is the one with the shortest history. Pew Research Center's Americans and AI 2026 survey, fielded February 17 to 23, 2026 among 5,119 US adults, found that 49% now say they have used an AI chatbot at all, up from roughly a third the year before. Of those users, 42% say they use a chatbot specifically to search for information, and 60% of US adults say they now read the summary a chatbot or search engine writes at the top of the results page.

The scale claims around this surface change quickly and should be read carefully. DemandSage's compilation of ChatGPT's weekly active users shows growth from 50 million in January 2023 to 300 million in December 2024, 800 million in October 2025, and 900 million in February 2026. A further claim, that ChatGPT crossed a billion monthly active app users around May 2026, traces back to Sensor Tower app-analytics estimates reported by Reuters, not an official disclosure from OpenAI, and is best read as a third-party estimate rather than a hard count. Separately, DataReportal's own synthesis of generative AI platforms overall (combining ChatGPT's app and web numbers with OpenAI's self-reported weekly figure and an estimate of AI-agent users in China) arrives at more than a billion combined monthly users, while stating outright that 'an exact figure remains elusive.' Both estimates are well sourced. Neither is an audited census.

Strong, sourced evidence of a real, fast-growing surface. Not yet a census of it.

Fewer people are going looking at all

Underneath the platform-by-platform numbers is a behavioral shift that changes what 'being found' even means. Pew's research into how Americans encounter news found that 49% now say they mostly get news by happening to come across it rather than actively seeking it out, up from 39% when Pew first asked the question in 2019. The shift skews sharply by age: 52% of 18 to 29 year olds say they're more likely to encounter up-to-date information incidentally, compared with just 28% of adults 65 and older.

This matters more than it sounds. A business built around ranking for a search query is tuning for the moment someone goes looking. But if half your future customers are more likely to stumble across you inside a feed, a chat answer, or a friend's share than to type your category into a search bar, then showing up well when someone is actively searching only covers half the terrain. The other half is won or lost in moments nobody typed a query for.

Attention is starting to convert where it sits

Some of these surfaces aren't just where attention lives, they're where the purchase happens too. The Shopify blog, citing figures from Momentum Asia and a Statista forecast, reports that TikTok Shop generated 33 billion dollars in gross merchandise value in 2024, more than double the year before, on a platform with roughly 1.59 billion global monthly active users forecast for 2025. Those specific figures are secondhand, aggregated through Shopify rather than measured firsthand, so treat the scale as directional. The pattern behind it, short-video platforms becoming discovery-to-purchase surfaces rather than pure entertainment, is the part worth taking seriously.

None of this erases the oldest, best-documented friction point in commerce. Baymard Institute's meta-analysis of 50 independently published studies spanning 2006 to 2025 puts average cart abandonment at 70.22%. Wherever the attention originates, whichever surface earned the click, most of it still leaks out at the final step. Winning a new surface doesn't excuse a checkout that loses seven in ten shoppers who got that far.

Outside the US, the terrain is already further along

If you sell only in the US, it's tempting to treat this shift as gradual. The Reuters Institute's Digital News Report 2025, based on fieldwork across 47 countries, shows how much further along the same shift already is elsewhere. TikTok is a primary news source for 49% of people in Thailand, 48% in Malaysia, and 40% in Kenya. In Brazil, print news reach has collapsed from 50% in 2013 to just 10% today, and 9% of Brazilians already say they get news from an AI chatbot, a figure with no US equivalent yet in this evidence set.

This is a preview, not a foreign curiosity. The Reuters Institute data is one slice of a much larger 47-country dataset, and the four markets above are simply the ones verified this pass. Read them as an early signal of where behavior elsewhere may be heading, not a global census.

The idea is old. The speed is new.

None of this requires a new theory of how people decide what to trust. Procter & Gamble's A.G. Lafley described the First and Second Moments of Truth back in 2005, the idea that a shopper forms or revises an opinion of a brand at every discrete point of contact, not at one climactic moment. Google extended the idea in 2011 with the Zero Moment of Truth, describing the research shoppers do before they ever reach a store or a checkout, citing 88% of US customers researching online before buying. The underlying theory goes back further still, to Herbert Simon's 1971 observation that a wealth of information creates a poverty of attention, the founding idea behind treating attention itself as a scarce, allocatable resource.

What's new isn't the theory. It's the count of surfaces and the speed at which that count is changing. In 2005, the moments of truth mostly happened on a shelf, a TV ad, and a phone call. Today they happen across search, five or six social platforms, a podcast feed, a marketplace listing, and increasingly, a single paragraph a chatbot writes in response to a question you'll never see asked.

The one surface with no scoreboard yet

Every surface described above has an established, independent measurement tradition behind it. Search share has trackers. Platform reach has Pew and GWI-fed panels. National news habits have the Reuters Institute. Audio has Edison Research's probability-sampled diary panel. The AI-answer layer, chatbots and features like Google's AI Overviews sitting at the top of search results, has none of that yet. There is no equivalent of a search-share tracker or a standing reach panel that independently, verifiably counts how often a given business gets cited inside an answer one of these engines produces.

The closest thing academia has offered so far is a 2023 to 2024 paper from researchers at Princeton, Georgia Tech, and the Allen Institute for AI, which found that systematically adapting content for citation inside the answers a generative engine produces, a discipline the paper calls Generative Engine Optimization, can lift a source's visibility in those answers by up to 40%, with the effect varying by content domain. That's a genuinely useful, peer-reviewed finding about how to improve your odds. It is not a population-level measurement of who is winning that visibility today, in your market, against your competitors. Nobody has built that instrument yet, at least not one independently audited and standing the way Pew's panel or Edison's diary study is.

Every surface here has an independent scoreboard except the newest, fastest-growing one.

What this means for your business

Put together, the evidence describes a market where roughly half of US adults now use an AI chatbot at all, where about the same share now stumbles onto information rather than searching for it, where search itself is still the largest single surface but is shrinking at its edges, and where the newest and fastest-growing surface has no independent census counting who wins it. None of that is a reason to panic. It's a reason to stop assuming last year's channel mix still describes this year's market.

Given everything above, no single report, including this one, can tell you exactly where your specific customers' attention sits today without measuring your specific market. The instruments exist for search, social, audio, and news. They don't yet exist, independently and at scale, for the AI-answer layer. That gap is where primary measurement, reading the actual terrain rather than assuming it, earns its keep.

The evidence, in numbers

Key findings, dated and sourced

  • Monthly search-engine usage among online adults fell to 80.3% in the past-month window, a decline of roughly 210 basis points year over year, the lowest level in recent tracking

    established DataReportal / We Are Social / Meltwater, drawing on GWI data, Digital 2026 Global Overview Report (2025-10)

  • Generative AI platforms are estimated to have surpassed 1 billion combined monthly users, per DataReportal's synthesis of several partial proxies (ChatGPT mobile app 550M+ MAU, ChatGPT web roughly 500M unique devices as of Aug 2025, OpenAI's self-reported 800M weekly users in Oct 2025, plus roughly 250M China AI-agent users). DataReportal itself states 'an exact figure remains elusive,' a directional, multi-source estimate, not an audited census

    emerging DataReportal / We Are Social / Meltwater, drawing on GWI data, Digital 2026 Global Overview Report (2025-10)

  • Online adults spend an average of 2.5+ hours daily across social networks and video platforms combined (18h36m weekly), and the average social media user is active on 6.75 different platforms each month

    established DataReportal / We Are Social / Meltwater / GWI, Digital 2026 Global Overview Report (2025-10)

  • About half of US adults (49%) now say they have used an AI chatbot, up from roughly a third in 2024; 42% use chatbots specifically for 'searching for information,' and 60% report reading the summary an engine writes at the top of search results

    established Pew Research Center, Americans and AI 2026: Chatbots, Smart Devices and Views on Impact (surveyed Feb 17-23 2026, n=5,119 US adults via the American Trends Panel, published June 17 2026) (2026-02)

  • Half of US adults (49%) say they mostly get news by happening to come across it rather than actively seeking it out, up from 39% when Pew first asked the question in 2019; younger adults (18-29) encounter news incidentally far more than seniors 65+ (52% vs 28% for up-to-date information)

    established Pew Research Center, What types of news do Americans seek out or happen to come across? (surveyed Dec 8-14 2025, n=3,560 US adults via the American Trends Panel, published Apr 20 2026) (2025-12)

  • Platform reach among US adults is uneven and shifting: YouTube 84%, Facebook 71%, Instagram 50%, Pinterest 37%, TikTok 32%, Snapchat 26%, LinkedIn 25%, WhatsApp 21%, X 8%

    established Pew Research Center, Social Media Fact Sheet (surveyed Feb 5-Jun 18 2025, n=5,022 US adults, address-based sampling and multimode protocol) (2025-06)

  • TikTok has become a primary news surface in several national markets outside the US (49% Thailand, 48% Malaysia, 40% Kenya), AI chatbots already reach 9% of Brazilians for news, and print news in Brazil has collapsed to 10% reach from 50% in 2013

    established Reuters Institute for the Study of Journalism, University of Oxford, Digital News Report 2025 (fieldwork/analysis by YouGov, 47 countries) (2025-06-17)

  • Generative Engine Optimization (GEO) techniques, systematically adapting content for citation inside the answers a generative engine produces, can boost a source's visibility in those responses by up to 40%, with effectiveness varying substantially by content domain

    established Princeton University / Georgia Tech / Allen Institute for AI (KDD 2024), Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan, Deshpande, "GEO: Generative Engine Optimization" (arXiv:2311.09735, revised 2024-06-28) (2023-11-16)

  • The average documented ecommerce cart abandonment rate across 50 independently published studies (spanning 2006-2025) is 70.22%

    established Baymard Institute, Cart Abandonment Rate Statistics (meta-analysis of 50 studies) (2025-09-22)

  • ChatGPT weekly active users grew from 50 million (Jan 2023) to 300 million (Dec 2024) to 800 million (Oct 2025) to 900 million (Feb 2026). ChatGPT reportedly crossed 1 billion global monthly active app users around May 2026, per Sensor Tower app-analytics estimates as reported by Reuters, the fastest consumer app to reach that milestone. This MAU figure is a third-party analytics estimate relayed through a vendor-stats aggregator, not an OpenAI official disclosure

    emerging DemandSage (compiling OpenAI-disclosed WAU series plus Sensor Tower estimates reported by Reuters for the MAU milestone), ChatGPT user-growth figures as compiled by DemandSage (2026-05)

  • 81% of Americans age 12+ (an estimated 233 million people) listened to online audio in the past month, podcast monthly reach stands at 58% (167 million), and YouTube monthly reach is 84%, audio and video-audio hybrids remain a large, stable attention surface distinct from text search and social feeds

    established Edison Research, The Infinite Dial 2026 (n=2,050 age 12+, fielded January 2026, SSRS Opinion Panel + probability sampling) (2026-03-12)

  • TikTok Shop generated $33 billion in gross merchandise value in 2024, more than doubling year over year, while the platform reached roughly 1.59 billion global monthly active users in 2025, evidence that short-video platforms are becoming direct commerce/discovery surfaces, not just entertainment. The GMV figure traces through a Momentum Asia press release and the MAU figure through a Statista forecast, both aggregated secondhand by the Shopify blog rather than measured firsthand by Shopify

    emerging Shopify blog, citing Momentum Asia (GMV) and Statista forecast data (MAU), TikTok Statistics: Users, Growth, and Shopping Data (2025)

  • The 'Moment of Truth' model, the idea that a customer forms or revises a brand impression at each discrete point of interaction with a brand rather than at one single decisive moment, was coined by P&G's A.G. Lafley in 2005 (First and Second Moments of Truth) and extended by Google in 2011 with the Zero Moment of Truth, describing pre-purchase research spread across many surfaces (Google cited 88% of US customers researching online before buying)

    established Procter & Gamble (Lafley, 2005) / Google (ZMOT, 2011), Moment of truth (marketing), Wikipedia concept summary

  • Attention economics treats human attention as a scarce, allocatable resource to which economic theory can be applied, the foundational academic framing, tracing to Herbert Simon's 1971 observation that 'a wealth of information creates a poverty of attention,' underlying any attempt to map where a population's finite attention is currently allocated

    established Herbert A. Simon / attention-economics literature, Attention economy, Wikipedia concept summary (1971)

Learning outcomes

What this study teaches

  1. Search still matters (80.3% monthly reach) but it is shrinking at the edges. Treat it as the largest surface you compete on, not the only one.
  2. Half of US adults already use an AI chatbot, and a similar share now finds information by accident rather than by searching, so a strategy built only around ranking for queries is covering roughly half the terrain.
  3. Platform reach is uneven and specific (YouTube 84%, Facebook 71%, TikTok 32%, and so on). Guessing where your customers spend time is riskier than checking.
  4. The AI-answer layer has no independent scoreboard yet, so any claim about 'AI visibility,' from any source including this one, should be read as directional until it is measured against your own market.
  5. Attention that converts on a new surface still loses most of it at checkout (70.22% average cart abandonment). Winning a new surface without fixing the last step does not move revenue on its own.

Honest limits

What this does not yet settle

  • No independent, audited census of the AI-answer layer exists yet. Search share has trackers, social platform reach has Pew and GWI, national news habits have the Reuters Institute, but there is no equivalent third-party-verified measurement of how often a brand is actually cited inside ChatGPT, AI Overviews, Perplexity, Gemini, or Copilot answers. This is precisely the gap primary capture (a Corpus, Search Surface Optimization, and a Machine-Readiness Score) is designed to fill, and no finding above should be read as a substitute for it.
  • No single cross-surface 'share of attention' index unifies search, social, short video, CTV, audio, messaging, marketplaces, maps, and AI answers on comparable methodology. Each source in this study measures its own surface with its own sample frame and cadence (Pew is a US probability panel, DataReportal is aggregated multi-vendor traffic and panel data, the Reuters Institute is a 47-country YouGov online panel, Edison is a US diary and phone panel), so cross-surface comparisons here are directional, not statistically equatable.
  • The behavioral-science literature on multi-touchpoint research journeys (the Moment of Truth model, 2005) predates generative AI by nearly two decades. The GEO paper (2023-24) is the closest academic anchor for the AI-answer era, but it studies a content-tuning technique, not population behavior, so there is not yet a mature body of research measuring how consumers actually allocate attention to the AI-answer layer relative to other surfaces.
  • The non-US evidence gathered here is limited to a handful of country cuts (Reuters Institute figures for Thailand, Malaysia, Kenya, and Brazil); a full pull of the Reuters Institute's complete 47-country dataset, and non-US equivalents to Pew and Edison, would be needed before generalizing terrain claims globally.
  • Several planned lines of inquiry for this pass, including direct AI Overviews click-through-rate studies and Google's own AI Mode scale disclosures, could not be independently verified and were deliberately left out rather than asserted from memory.

This is a synthesis of dated, attributed evidence, not a census. The AI-answer layer in particular has no independent, Nielsen-grade measurement yet, so readings of it are directional and named as a frontier, never presented as settled.

Straight answers

Frequently asked questions

Is search actually declining, or is this overstated?

Search is still the largest single surface, but it is shrinking at the edges. DataReportal's Digital 2026 Global Overview Report puts monthly search engine use among online adults at 80.3% as of October 2025, down roughly 210 basis points year over year and the lowest level the report has tracked. Eighty percent is still a large majority, so this is a story about search no longer being the automatic first stop it once was, not a story about search disappearing.

How many people are actually using AI chatbots to search for things?

Pew Research Center found that 49% of US adults have now used an AI chatbot at all, up from roughly a third the year before, and that 42% of chatbot users say they use one specifically to search for information. Pew also found 60% of US adults now say they read the summary a chatbot or search engine writes at the top of a results page, which means a meaningful share of searching no longer produces a list of links at all.

Can I actually measure how often my business gets cited inside ChatGPT or an AI Overview?

Not yet, at least not with an independent, audited instrument. The study notes that search share has trackers, social platform reach has Pew and GWI-fed panels, and national news habits have the Reuters Institute, but there is no equivalent third-party-verified measurement of how often a brand is cited inside ChatGPT, AI Overviews, Perplexity, Gemini, or Copilot answers. DataReportal itself describes its own estimate of generative AI platform users as one where 'an exact figure remains elusive,' stitched together from partial counts rather than one audited number. That gap is the biggest limitation in this evidence set, and it is exactly what primary market measurement is meant to close.

Is it proven that tuning content for AI answers actually works?

There is early, peer-reviewed evidence that it can help, but it is not the same as knowing who is winning visibility in your specific market today. A 2023 to 2024 paper from researchers at Princeton, Georgia Tech, and the Allen Institute for AI found that systematically adapting content for citation inside generative-engine answers, a technique the paper calls Generative Engine Optimization, lifted a source's visibility in those answers by up to 40%, with the effect varying by content domain. That is a useful finding about how to improve your odds. It is not a population-level measurement of who is actually getting cited in your market against your competitors, since no independently audited instrument for that exists yet.

If attention is spreading across so many surfaces, what should a small business actually do?

Stop assuming last year's channel mix still describes this year's market, and treat search as the largest surface you compete on rather than the only one. The evidence shows platform reach is uneven and specific, for example Pew found YouTube reaches 84% of US adults versus TikTok at 32%, so guessing where your customers spend time is riskier than checking. It's also worth remembering that winning a new surface doesn't fix conversion on its own: Baymard Institute's meta-analysis of 50 studies puts average cart abandonment at 70.22%, so attention that converts on a new surface can still be lost at the final step.

Provenance

References

  1. DataReportal (We Are Social / Meltwater, GWI data), Digital 2026 Global Overview Report, October 2025 https://datareportal.com/reports/digital-2026-global-overview-report
  2. Pew Research Center, Americans and AI 2026: Chatbots, Smart Devices and Views on Impact, June 17 2026 https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/
  3. Pew Research Center, What types of news do Americans seek out or happen to come across?, April 20 2026 https://www.pewresearch.org/short-reads/2026/04/20/what-types-of-news-do-americans-seek-out-or-happen-to-come-across/
  4. Pew Research Center, Social Media Fact Sheet, 2025 https://www.pewresearch.org/internet/fact-sheet/social-media/
  5. Reuters Institute for the Study of Journalism, University of Oxford, Digital News Report 2025 https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2025
  6. Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan, Deshpande, "GEO: Generative Engine Optimization," arXiv:2311.09735 (KDD 2024) https://arxiv.org/abs/2311.09735
  7. Baymard Institute, Cart Abandonment Rate Statistics https://baymard.com/lists/cart-abandonment-rate
  8. DemandSage, ChatGPT Statistics https://www.demandsage.com/chatgpt-statistics/
  9. Edison Research, The Infinite Dial 2026 https://www.edisonresearch.com/the-infinite-dial-2026/
  10. Shopify blog, TikTok Statistics: Users, Growth, and Shopping Data, citing Momentum Asia and Statista https://www.shopify.com/blog/tiktok-statistics
  11. Wikipedia, Moment of truth (marketing) https://en.wikipedia.org/wiki/Moment_of_truth_(marketing)
  12. Wikipedia, Attention economy https://en.wikipedia.org/wiki/Attention_economy

Every measured figure is dated to its capture and tagged with an evidence tier. Every cited work is real and locatable. Where an engine could not be captured this round, it is named as uncaptured, not estimated. Small-sample readings are labelled as directional.

Read your own market's terrain

You do not have to guess where your customers' attention actually sits, or take any single report's word for it, including this one. Raveneye Global builds a Corpus for your specific market: a map of where the people who would buy from you are actually spending their attention across search, social, video, audio, and the AI-answer layer, read against your competitors in that same terrain. From there we tune for the surfaces that matter most to you, what we call Search Surface Optimization, and track the result with a Machine-Readiness Score that measures whether you are gaining ground or losing it.