AI has not replaced search, but it has changed the unit of competition from “ranking a page” toward “being selected, cited or recommended inside an answer.” By 2026, AI summaries are mainstream in search, AI-referral traffic is growing rapidly, and high-intent AI-referred shoppers can convert unusually well. At the same time, publishers face fewer outbound clicks and the web is becoming more saturated with AI-authored content. The result is a fragmented discovery economy in which brands must optimise not only for traffic, but for machine comprehension, source authority, recommendation share and downstream conversion.
Research question and scope
Core question: How is AI changing the way people find information, products and providers, and what does that mean for search, publishers, brands and customer acquisition?
Scope: Global AI assistants and search surfaces with deeper evidence from the US where behavioural measurement is strongest. Covers search summaries, referral traffic, content/citation behaviour, retail conversion and machine visibility.
Method: Triangulates browsing-panel research, behavioural web analytics, Adobe transaction/traffic data, global consumer surveys and platform product changes. Percentage growth is always interpreted against base size where possible.
1. Discovery is becoming answer-led
Six in ten US adults reported in February 2026 that they read AI summaries at the top of search-engine results, rising to 72% among adults aged 18–29. S1 Similarweb reports that more than 40% of US searches now trigger an AI Overview, while Google remains by far the world’s largest website and search platform. S2S3 This is not a simple migration from Google to chatbots; AI is being incorporated into both incumbent search and standalone assistants.
2. The click economy is weakening for some information journeys
Pew’s behavioural analysis of 900 US adults found that when a Google result page contained an AI summary, users clicked a traditional search result in 8% of visits versus 15% without an AI summary. A cited link inside the summary was clicked in only 1% of visits. Users were also more likely to end the browsing session after a page with an AI summary. S4
Publisher-level data point in the same direction. Similarweb reports organic search traffic to news sites down 26% since Google launched AI Overviews, while ChatGPT referrals to publishers increased sharply from a much smaller base. S5 The structural tension is clear: AI surfaces can increase source exposure while reducing the need to visit the source.
| Observed change | Evidence | Implication |
|---|---|---|
| AI summaries mainstream | 60% of US adults say they read them. S1 | Answer consumption is no longer niche. |
| Fewer outbound search clicks | 8% click-through with AI summary vs 15% without in Pew browsing sample. S4 | Traffic-based publisher economics face pressure. |
| AI referrals growing | Similarweb: average 770.7m monthly AI referral visits globally Jun 2025–May 2026, +117.4% YoY. S6 | New referral channel is becoming measurable. |
| Retail traffic quality improving | Adobe: Mar 2026 AI-referred retail visitors converted 42% better than non-AI traffic. S7 | AI can compress research and increase purchase readiness. |
| Content supply changing | Pew: ~10% of sampled English webpages in Jul 2026 show significant signs of AI authorship; >one-third among pages published after ChatGPT. S8 | Authority and provenance become more important amid content abundance. |
3. AI referrals are small relative to search—but strategically important
Similarweb’s 2026 cross-industry analysis estimates 770.7 million average monthly visits referred by AI platforms worldwide from June 2025 through May 2026, up 117.4% year over year. S6 Against Google’s approximately 87.5 billion monthly visits in May 2026, the absolute scale remains far smaller. S3 The correct strategic conclusion is therefore not “AI search has replaced Google,” but “a fast-growing referral and influence layer is emerging inside a still search-dominated market.”
4. AI is creating higher-intent visits in commerce
Adobe Digital Insights, drawing on very large-scale digital transaction/traffic data, reports that traffic from AI sources to US retail sites grew 393% year over year in Q1 2026. By March, AI-referred visitors converted 42% better than non-AI traffic, spent 48% longer on site and viewed more pages. S7 By July, Adobe reported AI-referred retail traffic up 62% year over year and conversion 60% higher than non-AI traffic, marking the eleventh consecutive month in which AI conversion had outperformed. S9
The interpretation requires caution: AI referral volumes remain modest; US retail is not representative of all categories; and users who choose an AI assistant may already be more research-oriented. Even so, the persistence of the conversion premium suggests AI may increasingly function as pre-qualification rather than merely awareness.
Sources and notes
- Pew Research Center, Americans and AI 2026 (17 Jun 2026). [Source](https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/)) — US survey on AI summaries and chatbot use.
- Similarweb, 2026 Generative AI Landscape Report. [Source](https://www.similarweb.com/corp/reports/2026-generative-ai-landscape/)) — Behavioural platform/search data; vendor research.
- Similarweb, Top 100 websites worldwide — May 2026. [Source](https://www.similarweb.com/blog/research/market-research/most-visited-websites/)) — Absolute scale context for Google and major platforms.
- Pew Research Center, Google users click less with AI summaries (22 Jul 2025). [Source](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/)) — Behavioural browsing study of 900 US adults.
- Similarweb, Impact of Generative AI on Publishers. [Source](https://www.similarweb.com/corp/reports/generative-ai-publishers/)) — Publisher traffic and ChatGPT referral evidence.
- Similarweb, AI Referral Traffic by Industry: 2026 Data. [Source](https://aisearch.similarweb.com/blog/ai-referral-traffic-by-industry/)) — Global cross-industry AI referral estimates.
- Adobe Digital Insights, Q2 2026 AI-sourced traffic report. [Source](https://business.adobe.com/resources/sdk/.2026-q2-ai-traffic-report/q2-2026-adi-ai-sourced-traffic-insights.pdf)) — US retail behavioural and conversion data.
- Pew Research Center, How Much of the Internet Is Written With AI? (20 Aug 2026). [Source](https://www.pewresearch.org/data-labs/2026/08/20/how-much-of-the-internet-is-written-with-ai/)) — 490,000-page Common Crawl sample; probabilistic AI-authorship detection.
- Adobe, AI reshaping travel planning and July retail update (Aug 2026). [Source](https://business.adobe.com/blog/us-consumers-are-embracing-llms-to-make-travel-plans)) — Latest cross-category and July 2026 retail update.
8i TECHNOLOGIES
The New Discovery Economy
How AI search, answers and assistants are changing traffic, brand visibility, customer journeys and the economics of being found
Reference date: 15 September 2026
See what matters next.
| Executive answer AI has not replaced search, but it has changed the unit of competition from “ranking a page” toward “being selected, cited or recommended inside an answer.” By 2026, AI summaries are mainstream in search, AI-referral traffic is growing rapidly, and high-intent AI-referred shoppers can convert unusually well. At the same time, publishers face fewer outbound clicks and the web is becoming more saturated with AI-authored content. The result is a fragmented discovery economy in which brands must optimise not only for traffic, but for machine comprehension, source authority, recommendation share and downstream conversion. |
|---|
Research question and scope
Core question: How is AI changing the way people find information, products and providers, and what does that mean for search, publishers, brands and customer acquisition?
Scope: Global AI assistants and search surfaces with deeper evidence from the US where behavioural measurement is strongest. Covers search summaries, referral traffic, content/citation behaviour, retail conversion and machine visibility.
Method: Triangulates browsing-panel research, behavioural web analytics, Adobe transaction/traffic data, global consumer surveys and platform product changes. Percentage growth is always interpreted against base size where possible.
1. Discovery is becoming answer-led
Six in ten US adults reported in February 2026 that they read AI summaries at the top of search-engine results, rising to 72% among adults aged 18–29. [S1] Similarweb reports that more than 40% of US searches now trigger an AI Overview, while Google remains by far the world’s largest website and search platform. [S2][S3] This is not a simple migration from Google to chatbots; AI is being incorporated into both incumbent search and standalone assistants.
2. The click economy is weakening for some information journeys
Pew’s behavioural analysis of 900 US adults found that when a Google result page contained an AI summary, users clicked a traditional search result in 8% of visits versus 15% without an AI summary. A cited link inside the summary was clicked in only 1% of visits. Users were also more likely to end the browsing session after a page with an AI summary. [S4]
Publisher-level data point in the same direction. Similarweb reports organic search traffic to news sites down 26% since Google launched AI Overviews, while ChatGPT referrals to publishers increased sharply from a much smaller base. [S5] The structural tension is clear: AI surfaces can increase source exposure while reducing the need to visit the source.
| Observed change | Evidence | Implication |
|---|---|---|
| AI summaries mainstream | 60% of US adults say they read them. [S1] | Answer consumption is no longer niche. |
| Fewer outbound search clicks | 8% click-through with AI summary vs 15% without in Pew browsing sample. [S4] | Traffic-based publisher economics face pressure. |
| AI referrals growing | Similarweb: average 770.7m monthly AI referral visits globally Jun 2025–May 2026, +117.4% YoY. [S6] | New referral channel is becoming measurable. |
| Retail traffic quality improving | Adobe: Mar 2026 AI-referred retail visitors converted 42% better than non-AI traffic. [S7] | AI can compress research and increase purchase readiness. |
| Content supply changing | Pew: ~10% of sampled English webpages in Jul 2026 show significant signs of AI authorship; >one-third among pages published after ChatGPT. [S8] | Authority and provenance become more important amid content abundance. |
3. AI referrals are small relative to search—but strategically important
Similarweb’s 2026 cross-industry analysis estimates 770.7 million average monthly visits referred by AI platforms worldwide from June 2025 through May 2026, up 117.4% year over year. [S6] Against Google’s approximately 87.5 billion monthly visits in May 2026, the absolute scale remains far smaller. [S3] The correct strategic conclusion is therefore not “AI search has replaced Google,” but “a fast-growing referral and influence layer is emerging inside a still search-dominated market.”
4. AI is creating higher-intent visits in commerce
Adobe Digital Insights, drawing on very large-scale digital transaction/traffic data, reports that traffic from AI sources to US retail sites grew 393% year over year in Q1 2026. By March, AI-referred visitors converted 42% better than non-AI traffic, spent 48% longer on site and viewed more pages. [S7] By July, Adobe reported AI-referred retail traffic up 62% year over year and conversion 60% higher than non-AI traffic, marking the eleventh consecutive month in which AI conversion had outperformed. [S9]
The interpretation requires caution: AI referral volumes remain modest; US retail is not representative of all categories; and users who choose an AI assistant may already be more research-oriented. Even so, the persistence of the conversion premium suggests AI may increasingly function as pre-qualification rather than merely awareness.
5. Product discovery is moving inside assistants
OpenAI expanded ChatGPT product discovery in March 2026 with richer comparison and up-to-date product information using the Agentic Commerce Protocol, demonstrating how a general AI interface can absorb functions historically distributed across search, reviews, retailer sites and comparison tools. [S10] Salesforce’s 2026 commerce research reports agentic search as the first stage of the shopping journey growing 200% year over year and combines survey evidence with behavioural data across more than 1.5 billion shoppers. [S11]
6. What determines AI visibility?
Traditional search optimisation remains relevant because AI systems often retrieve from the same public web, but the objective is broader. An AI system must be able to identify what an entity is, trust its information, resolve attributes, compare it against alternatives and answer a user’s specific intent. This shifts optimisation toward structured meaning and evidence.
| Visibility factor | Human-search role | AI-discovery role |
|---|---|---|
| Entity clarity | Helps relevance and navigation | Allows models to resolve product/company identity and relationships |
| Structured data | Supports rich results | Provides machine-readable attributes and availability |
| Authoritative evidence | Builds SEO authority | Supports citation/recommendation confidence |
| Freshness | Important for prices/news | Critical where AI claims current facts or availability |
| Third-party mentions | Backlinks/reputation | Cross-source corroboration and recommendation presence |
| Specificity | Long-tail ranking | Answering detailed intent and comparison prompts |
| Accessible content | Indexability | Retrievability by crawlers, agents and data feeds |
7. The web itself is filling with AI-generated material
Pew’s August 2026 analysis sampled 490,000 English-language webpages across 49 Common Crawl snapshots from 2021 to July 2026. It found significant signs of AI authorship in 10% of all pages in the July 2026 snapshot and more than one-third of pages published after ChatGPT’s release. [S8] Detection is probabilistic and does not establish whether content is poor, but it demonstrates a supply shock: producing plausible text is no longer scarce.
| 8i interpretation As content abundance rises, the scarce resources shift toward original evidence, trusted identity, expert judgment, proprietary data, demonstrated experience and continuously maintained facts. That supports the strategic logic of 8i Intelligence: referenceable intelligence should become more valuable as generic content becomes cheaper. |
|---|
8. Winners and losers in the discovery transition
| Actor | Potential advantage | Potential risk |
|---|---|---|
| Search/AI platforms | Own intent and answer interface | Trust/regulatory burden; source ecosystem deterioration |
| Brands/retailers | Higher-intent AI referrals; direct product data integrations | Loss of control over presentation and ranking |
| Publishers | Citation/brand exposure; new referral sources | Zero-click answers and weaker pageview economics |
| Marketplaces/comparison sites | Rich structured catalogues and transaction capability | AI may perform comparison directly |
| Niche experts | High authority in specific questions | Harder to monetise if answers consume expertise without visits |
| Customers | Less research friction, personalised synthesis | Opaque recommendation criteria, errors and reduced source awareness |
9. Measuring the new discovery funnel
- Presence: does the brand/entity appear in AI answers for priority intents?
- Share of recommendation: how often is it selected relative to competitors?
- Source/citation share: which domains are used as evidence?
- AI referral volume and quality: conversion, engagement, order value, returns and retention.
- Answer accuracy: are attributes, prices, policies and claims represented correctly?
- Machine-readiness: can agents retrieve structured, current and permissioned information?
- Economic impact: does AI discovery add incremental customers or merely reclassify existing traffic?
10. 8i assessment
| 8i view — ACT NOW, MEASURE CAREFULLY AI-mediated discovery is already strategically material, even though traditional search remains much larger. Organisations should treat it as a new distribution surface, not a replacement channel. The highest-value work now is measurement, machine-readiness and authoritative information—not speculative “GEO hacks.” |
|---|
Source register
Sources are prioritised toward primary institutions, transparent methodology, observable behavioural data, large-sample surveys and current production evidence. Vendor evidence is identified as such and used cautiously where it provides proprietary usage data or live implementation detail.
[S1] Pew Research Center, Americans and AI 2026 (17 Jun 2026). Source — US survey on AI summaries and chatbot use.
[S2] Similarweb, 2026 Generative AI Landscape Report. Source — Behavioural platform/search data; vendor research.
[S3] Similarweb, Top 100 websites worldwide — May 2026. Source — Absolute scale context for Google and major platforms.
[S4] Pew Research Center, Google users click less with AI summaries (22 Jul 2025). Source — Behavioural browsing study of 900 US adults.
[S5] Similarweb, Impact of Generative AI on Publishers. Source — Publisher traffic and ChatGPT referral evidence.
[S6] Similarweb, AI Referral Traffic by Industry: 2026 Data. Source — Global cross-industry AI referral estimates.
[S7] Adobe Digital Insights, Q2 2026 AI-sourced traffic report. Source — US retail behavioural and conversion data.
[S8] Pew Research Center, How Much of the Internet Is Written With AI? (20 Aug 2026). Source — 490,000-page Common Crawl sample; probabilistic AI-authorship detection.
[S9] Adobe, AI reshaping travel planning and July retail update (Aug 2026). Source — Latest cross-category and July 2026 retail update.
[S10] OpenAI, Powering Product Discovery in ChatGPT (24 Mar 2026). Source — Platform product evidence; vendor source.
[S11] Salesforce, State of Commerce / Agentic Search (28 Jul 2026). Source — 4,690 consumers, 3,450 commerce professionals + >1.5bn shopper behavioural data.
[S12] Adobe, AI traffic surge / machine-readable retail sites (Apr 2026). Source — Retail machine-readiness and engagement evidence.