The Shift
No permission needed. Free to reproduce with attribution under CC BY 4.0.
AIGNCI Digital. "The Shift: August 2026 Edition." AIGNCI Digital Market Brief, August 2026. https://www.aignci.com/the-shift
Buyers no longer start with ten blue links.
They ask an engine, and the engine answers with a shortlist. Either a company is in that answer or a competitor is. The 17 figures below are why, and each one states its publisher and its date so it can be checked rather than taken on faith.
weekly ChatGPT users, up from 400 million one year earlier. The app passed one billion monthly users in June 2026.
How this can be misread: Weekly and monthly user counts are different measures and should not be compared with each other.
of tracked Google searches now return an AI Overview, a 58 percent increase in a single year.
How this can be misread: The 58 percent is relative growth in the share, not 58 percentage points. Coverage also varies sharply by query category.
GenAI chatbots now rank first in influencing B2B vendor shortlists, ahead of review sites and vendor websites.
AI-referred visitors convert 42 percent better than all other traffic. A year earlier the same cohort converted 38 percent worse.
How this can be misread: The reversal is the finding. A single year's figure in isolation says much less than the direction of travel.
The B2B buyer has moved
Vendor selection now begins inside an assistant, and the traffic that arrives from one behaves differently to search traffic.
of B2B technology buyers use AI search in their initial vendor research, up from 17 percent in 2023 — a 3.4-fold rise in under three years.
of B2B buyers now use AI tools such as ChatGPT and Perplexity somewhere in their research process.
the conversion rate of standard organic traffic, measured across more than 500 high-value B2B topics.
Basis: 500+ high-value B2B topics
of Fortune 500 companies have ChatGPT somewhere in their workflows.
How this can be misread: Presence in a workflow is not procurement influence. This measures how widely the tool is in use, not what it decides.
The consumer has moved
The same behavior appears in retail, where the effect is large enough to be visible in single-day sales figures.
ChatGPT's share as a product research tool — a fifteen-fold rise in two years.
year-over-year growth in AI-referred traffic to US retail sites in the first quarter of 2026, following a 693 percent holiday surge.
Basis: Over one trillion US retail visits
How this can be misread: Growth rates this large start from a very small base. The absolute share of retail traffic remains modest.
in online Black Friday sales influenced by AI in a single day.
of consumers plan to use generative AI to shop in 2026, and 85 percent of those who already have say it improved the experience.
How this can be misread: Stated intent consistently overstates behavior. The 85 percent satisfaction figure among actual users is the sturdier half.
The gap most companies have not seen yet
Ranking and being cited have come apart, and almost nobody is measuring the difference.
the collapse, in under a year, in overlap between the organic top ten and the pages AI answers actually cite. Ranking first on Google no longer means being in the AI answer.
How this can be misread: The range is wide because method and vertical change the result substantially. See the correction below — an earlier edition printed a narrower range from a single source.
of Fortune 500 companies track their AI search performance at all, so most of the market cannot yet see which answers it is missing.
Where AI citations actually come from
New in this edition, and the most directly actionable group: the engines overwhelmingly cite earned and third-party sources, and they read far more than they send back.
of all AI citations come from earned media — journalism, academic, government, encyclopedic and other third-party sources. Paid and advertorial content accounts for 0.3 percent.
Basis: More than 25 million links
the click rate on a Google result with an AI summary present versus without one. Just 1 percent of visits produce a click on a source cited inside the summary.
Basis: 68,879 searches
How this can be misread: This is the figure that argues against treating AI citation as a traffic channel. Being cited is a reputational position, not a click source.
of AI crawler requests are search-purpose — the kind that can return a citation. The engines read far more than they send back.
How this can be misread: Window-specific. Crawler purpose ratios move with model release cycles and should not be averaged across periods.
What we got wrong, and what it says now.
A figure that is quietly changed between editions is not a figure anyone should rely on. When we revise one, the previous reading, the new reading and the reason all stay on the page.
- As printed
- The citation-overlap collapse was printed as 75 percent falling to a range of 17 to 38 percent, sourced to a single vendor study.
- As corrected
- It now reads 76 percent falling to a range of 17 to 54 percent, sourced to three separate bodies.
- Why
- A wider read of the same period puts the floor and ceiling further apart depending on method and vertical. We print the wider range because it is the honest one, and we broadened the sourcing from one vendor to three. The direction of the finding is unchanged.
Every source, and how independent it is.
Research in this category is published largely by companies selling into it. That does not make it wrong, and on several questions it is the only data that exists — but a reader deserves to know which findings come from a disinterested party and which come from a firm with a product to sell. Each figure above carries its tier. Here is what the three tiers mean.
Independent research
5 sourcesSurvey organizations, newsrooms, infrastructure operators and academic bodies with no product to sell in this category.
Cloudflare Radar · Muck Rack Generative Pulse · PYMNTS Intelligence · Pew Research Center · Reuters
Platform disclosure
5 sourcesFigures published by the companies operating the engines or measuring the traffic. Authoritative on their own systems, and self-interested about the size of them.
Adobe · Adobe Digital Insights · Capital One Shopping and Adobe · OpenAI · OpenAI, reported by Reuters
Vendor research
7 sourcesStudies published by firms that sell software or services in the search and answer-optimization market. Frequently the only data that exists on a question, and worth reading with the seller's interest in view.
Ahrefs, ALM Corp and BrightEdge · AirOps · Averi · BrightEdge · G2 Buyer Behavior Report · Katalysts · Semrush
Figures are third-party published research, dated as reported by each source and cited so they can be checked. They are not AIGNCI measurements. Coverage shares and crawler ratios are window-specific and should not be averaged across periods.
Every edition stays online.
A data resource that overwrites itself cannot be checked. Each edition keeps its own dated URL and its figures as published, so a reader can see what changed between one edition and the next — and so can we.
August 2026
Current edition · dated archive · 17 figuresSeventeen figures on how buyers now find companies, each carrying its publisher and date. This edition widens and re-sources the citation-overlap finding printed in July, and adds three findings on where AI citations actually come from.
July 2026
Archived as published · 14 figuresThe first edition, establishing that buyers had moved to AI-mediated research. Its citation-overlap finding was corrected in the August edition and is preserved here as published.
What the numbers ask of a company.
Three findings on this page point the same direction. Citations come overwhelmingly from earned and third-party sources, ranking no longer predicts being cited, and almost nobody is measuring the difference. That is a sequence of problems, not a single one, and it is the reason we work in the order we do.
Can the machine get the page at all?
Bot rules, CDN policy, redirect chains and pages that only assemble inside a browser decide this before a word of content matters. When it fails, you are not in the answer at any price and nothing reports it.
Can it find the answer inside?
Prices, locations, hours and capabilities have to be in the served HTML, labeled, current, and not contradicted by your own other pages. When it fails, the engine takes the fact from whoever published it plainly.
Will it name you?
Settled largely off your website — one consistent entity across every property you own, and third parties describing you accurately where the engines read. The 84 percent earned-media figure above is this gate stated as data.
About this brief
How often is The Shift updated?
Are these AIGNCI's own measurements?
May I cite The Shift?
Why do some figures carry a warning about how they can be misread?
Why are sources labeled independent, platform or vendor?
What happens to figures that turn out to be wrong?
Tell us what you are building.
If the question is about discovery, reputation, a new site, or a market move, send the context. AIGNCI will tell you whether an Audit, a build, or a more focused engagement is the right starting point.