Blog · 25 September 2026 · By HeardOf

Does AI traffic convert better? The published conversion figures, with what each one counted

Does AI traffic convert better than organic search?

In ten of the eleven studies we read that print a conversion figure for AI-referred visitors beside a figure for organic search or another baseline, the AI figure is higher in the headline. The eleven differ in what they count, whose sites and which window, so the gaps cannot be lined up: qualified enquiries per visit at 312 IT firms (Opollo, 22 February 2026, historical), purchases per session at 94 online stores (180 Marketing, 23 February 2026, historical), sign-ups on 1,277 publisher domains in one month (Microsoft Clarity, 6 November 2025, historical) and the share of one company's sign-ups (Ahrefs, 16 June 2025, historical) are four units, and a 14.2% in the first is not larger or smaller than a 1.81% in the second.

Three things sit against the headline direction. First Page Sage's report of 1 September 2026, the table's one current row with a publisher's own figure, prints ten industries with each industry's overall conversion rate beside its ChatGPT conversion rate: hotels and resorts sit at 3.6% overall and 7.2% from ChatGPT, but manufacturing at 2.1% and 2.0% and financial services at 1.8% and 1.5%, under a key-findings line that says ChatGPT referrals "convert at higher rates than other marketing channels across every industry in the dataset". And the eleventh study runs the other way: Kaiser and Schulze's paper in Marketing Science, published online 21 April 2026, historical, on 973 ecommerce sites, finds ChatGPT referrals converting above paid social and below every other channel, with organic search 13% more likely to convert after regression adjustment, in its main model on 2 February to 2 August 2025; the paper calls its core finding robust in checks that include twelve months from 28 July 2024, though the gap to organic search specifically is not statistically significant in five of its specifications and reverses sign, not significantly, in a sixth. The journal's page answered us with a bot challenge, which we did not pass, so we read the Internet Archive's capture of 17 May 2026; Pixis's post, published 23 April 2026 and modified 18 September 2026 per its metadata, relays the same historical finding as "roughly 13%" below organic. WebFX's industry table of 18 March 2026, historical, has one row the other way, energy and utilities at 37.78% against 38.52%.

What is the AI traffic conversion rate in these studies?

There is no single one, because an AI traffic conversion rate is the share of visits referred by an AI assistant that end in whatever the site counts as a conversion, and each study counts something different. For a B2B software buyer the nearest figures are First Page Sage's: 3.3% of ChatGPT referral visits to its B2B SaaS clients converted, on its page of 1 September 2026 covering "more than 160 client companies" from May 2025 to September 2026, where a conversion is "as defined by that company"; its page of 3 April 2026, historical, covering "150+ client companies" from May 2025 to April 2026, prints 2.4% for the same cell, with Gemini at 2.2%, Claude at 1.9% and Perplexity at 1.9%. Neither page, as served, says why the two ChatGPT figures differ, and the September page prints manufacturing at 4.6% in one table and 2.0% in the next. Both pages say most companies in the dataset "invested in generative engine optimization", a service First Page Sage sells.

The other figures are each on their own unit, and the table carries their samples and windows. Opollo's 14.2% counts booked consultations, demo requests and validated contact submissions per visit, newsletter sign-ups excluded, against 2.8% for Google organic at the same 312 firms. Seer Interactive's 15.9% (3 June 2025, historical) is GA4 "key events (conversion rate)" for one client whose industry the page does not name, against 1.76% for Google organic. Microsoft Clarity's 1.66% is a sign-up rate for visits from LLMs on 1,277 publisher domains in one month, against 0.15% from search. 180 Marketing's 1.81% is purchases per session against 1.39% for non-branded organic, with the order value lower, $204 against $238, so revenue per session ends 10.3% higher, as the page gives it. Similarweb's 11.4% (17 September 2025, historical) is its own estimate of visits "converted to sales" in June 2025, against 9.3% for paid search and 5.3% for organic. Goodie's figure (31 May 2025, historical) is not a visit rate at all: leads that arrived from AI search closed at a 56.3% higher rate than leads from Google and Bing, across 117,432 leads at 13 companies.

Why do two reports print the same 14.2% and 2.8%?

We do not know. Opollo's benchmark report of 22 February 2026, historical, gives an AI conversion rate of 14.2% and a Google organic rate of 2.8% for 312 firms measured from January 2025 to January 2026. A page on superprompt.com dated 18 September 2025, historical, gives 14.2% and 2.8% for 12.3 million visits across 347 businesses from January to September 2025, and adds that its "conversion data was normalized across industries using standardized goal values". Neither page mentions the other on what we read. RankScience's post of 9 October 2025, historical, quotes the Superprompt pair with a link to it, and Pixis describes RankScience's page as a "separate analysis of 12 million site visits" that "landed on nearly the same numbers independently". What we can say is that two samples of different size and window are printed with an identical pair of figures, and that one of the three pages carrying the pair is a relay.

Is Adobe's 42% a conversion rate?

No, it is a lift, and it is a lift over non-AI traffic (all non-AI traffic to U.S. retail sites, as Digital Applied and Digital Commerce 360 describe it) rather than over Google. As eMarketer states it on 7 May 2026, historical, "AI-referred retail visitors converted 42% more than non-AI visitors in March 2026, per Adobe"; Digital Applied (24 June 2026, historical) and Cognizo (13 August 2026) print the same figure with March 2025 at 38% worse, and Digital Commerce 360 (19 August 2026) prints Adobe's July 2026 reading of 60% higher, "the 11th straight month", with revenue per visit 53% higher, on "more than 1 trillion visits to U.S. retail sites". Digital Applied's own text says the "better than Google" phrasing in its title is shorthand, since "Adobe's comparison is against non-AI traffic broadly". Adobe's blog gave no response to two requests from us on 25 September 2026, so every Adobe figure here is as one of those four pages relays it, and none of them, as served, prints a rate. Digital Applied also attributes to "Shopify Enterprise Blog, May 2026" a figure of AI sessions converting about 50% higher than organic search on product detail pages; the page as served carries no link to it, and we did not find Shopify's page.

Which of these pages measured anything themselves?

Of the sixteen rows in the table, twelve come from a page where the publisher says it measured or estimated the figure on its own sites, clients or panel, two come from pages that relay another party's figure, and two are pages with no conversion figure. Pixis's post is a compilation whose headline of "4-5x" its own text says "is grounded primarily in Opollo's cross-industry B2B benchmark"; Digital Applied's lists "Adobe, Shopify, trade press" as its sources; Cognizo's says "the figures in this roundup come from several independent datasets". Goodie's 2026 AI Search Traffic Report of 21 May 2026, historical, prints engagement rates and engagement time by source, about 67.8% and 58.5 seconds for AI sources against 61.8% and 44.2 seconds for Google organic over 5 February to 5 May 2026, states "We don't make pipeline or revenue claims" and points to a funnel update due in the third quarter of 2026; its conversion figure is on the 2025 page above. Similarweb's AI Search Stats 2026 page of 29 July 2026 covers market share, referral share, citation rate and ads, and has no conversion figure as served or as rendered.

What each publisher sells, from its own pages: First Page Sage invites readers to "learn more about our GEO services"; Goodie sells an AEO platform with a pricing page; Pixis sells Pixis Visibility and ends its post with a demo booking; Similarweb sells AI Search Intelligence; Digital Applied offers an "agentic SEO and AEO service"; Cognizo sells an AI-visibility platform with a pricing page; Opollo, Seer Interactive and RankScience list AEO or GEO services in their navigation, 180 Marketing lists AI SEO services, and WebFX's page carries a proposal link; Ahrefs' footer lists Brand Radar and an AI Visibility Checker; Superprompt links its post to its own analytics product; Microsoft Clarity is Microsoft's analytics product, "free forever" per its page. That describes the pages and judges none of the figures.

Does HeardOf measure conversion?

No. Our audit records which brands an answer names and which pages it cites, from our own systems, and not whether anyone clicked or bought; the section "So does ChatGPT or Perplexity drive more traffic?" in who the AI actually cites answers "we do not know" for the same reason, about our two-engine research run of 4 September 2026, and which engines we ask is in our post on which engines we check, as updated 23 September 2026. What this compilation adds for a buyer is the unit column: when a vendor says AI traffic converts at four or five times the rate of search, one of these rows is where the figure came from, and the row says what it counted. How many of those visits arrive in the first place is a separate compilation: how much traffic AI sends puts the published share, growth and click figures beside what each one measured.

How was this read?

Every page was fetched on 25 September 2026 between 00:06 and 00:33 UTC with curl and a Chrome user-agent and read in full. Goodie's 2026 report and Similarweb's AI Search Stats page were also rendered in Chrome 153 headless, Goodie at 00:17 and Similarweb at 00:19 UTC, before the negative about each; the rendered text of each has the same word count as the served page and no conversion figure. Dates are the ones printed on each page; historical means older than 90 days on the reading day, and the ages run from 24 days for First Page Sage's September report to 482 for Goodie's funnel study. One machine, one location.

What we could not do: Adobe's blog gave no response to two requests, so Adobe's figures are as the four pages above relay them; the Marketing Science page for Kaiser and Schulze answered with a Cloudflare challenge, which we did not pass, so its text was read from the Internet Archive's capture of 17 May 2026; Shopify's May 2026 page was not linked and we did not find it; Similarweb's AI traffic checker page answered 403 from CloudFront twice, and its State of Ecommerce 2025 report address redirects to the 2026 edition, whose landing page prints 23% for shopper journeys involving both AI and search with no date as served, so we did not use it; and the Visibility Labs address Pixis links redirects to 180 Marketing, which is the page we read. Both Similarweb pages carry "Summarize with AI" links addressed to assistants, Cognizo's page builds the same kind of links for ChatGPT, Claude and Perplexity, and WebFX's page carries one to ChatGPT; all were read as data and not followed. Everything this post says about the pages named carries a link and the day it was read; what it says about our own audit comes from our own systems and cannot be checked from here.

How much traffic does AI send? Published figures, with what each measured