AI and the future of search

AI traffic: do AI answers bring clicks and customers?

AI traffic in 2026: what Pew, Ahrefs, Seer and four other studies measured on clicks and conversion, and how to judge what AI answers bring your business.

Written by Mike Bastin30 October 2026

What AI answers change for your clicks

A buyer who asks an AI assistant for suppliers gets a short list before visiting a single website. For a company that sells abroad, that list is now part of the sales process, in every language. The practical question is how much AI traffic the answers send, and whether the visitors who arrive buy.

Eight published studies answer it, each with a different sample and period. Read side by side, they show a small share of visits, a large effect on organic clicks and a conversion rate that depends on the site. Each figure below comes from the study's own page, checked on 30 October 2026.

Fewer clicks when an AI answer appears

Pew Research tracked the browsing of 900 US adults for a month and counted what they clicked after a Google search.

In just 1% of visits to results pages with an AI summary, users clicked a link inside that summary.

Source: Pew Research, "Google users are less likely to click on links when an AI summary appears in the results", browsing data of 900 US adults, March 2025, published 22 July 2025

The answer itself satisfies many searches, so the sources named inside it receive a small share of the visits. Ahrefs measured the same effect on organic results across 300,000 keywords.

The presence of an AI Overview now correlates with a 58% lower average clickthrough rate for the page in position one, comparing 150,000 keywords with an AI Overview and 150,000 informational keywords without one, December 2023 against December 2025.

Source: Ahrefs, "Update: AI Overviews Reduce Clicks by 58%", Ryan Law, desktop Google Search Console data, 4 February 2026

Ahrefs reports a correlation on desktop searches, so treat 58% as the size of a pattern rather than a forecast for your own pages. Ranking first still matters, and it brings fewer clicks when an AI Overview sits above it.

Being cited changes the odds

Seer Interactive followed 53 brands through 2025 and split their Google results by whether an AI Overview cited the brand.

On informational queries, brands cited in an AI Overview earned a 2.07% organic click-through rate in 2025, against 0.94% for brands not cited. The data covers 53 brands and 5.47 million tracked queries.

Source: Seer Interactive, "AIO Impact on Google CTR: 2026 Update", informational queries, full-year 2025 average, 24 April 2026

Seer states that it cannot claim causation, and it adds that a falling click-through rate is not automatically bad news: a page that appears for more searches can show a lower rate while its visits hold steady. A citation gives your company a better position inside the answer, which is the reason the GEO measurement approach tracks citations alongside clicks.

How much AI traffic do assistants send?

Two large datasets put the volume in proportion. Kevin Indig analysed Similarweb estimates for 10,000 domains in each of the US and France.

AI assistants account for about 0.5% of total visits, in Similarweb monthly estimates for 10,000 domains in each of the US and France, from September 2023 to August 2026. The author notes that the samples do not represent the entire national web.

Source: Kevin Indig, "Organic traffic is down, but it's not all AI", Growth Memo, 26 October 2026

The SEO Works compared the referrals of 95 websites over twelve months and reached the same order of magnitude.

Organic search delivered 39 times more traffic than AI assistants across 95 websites, from September 2025 to August 2026.

Source: The SEO Works, "How Much Traffic Do AI Assistants Actually Send, And Does It Convert Better?", agency benchmark, September 2025 to August 2026

Google organic remains the main source of visits for the sites measured. The assistants add a small channel on top of it. Kevin Indig adds that hidden AI traffic is hard to detect, so the 0.5% describes what analytics can see.

Does AI traffic convert better?

Four studies measured conversion, and they give different answers. Marie Haynes analysed 100 sites and compared the 55 that track conversions in GA4.

ChatGPT visitors converted at 13.8% against 3.02% for Google organic visitors across the 55 of 100 sites tracking conversions in GA4 over 12 months. ChatGPT sessions amounted to 0.62% of the volume driven by Google organic.

Source: Marie Haynes, "GEO is a distraction", 21 October 2026

The SEO Works found a smaller gap on its sample of 63 sites, and a split result.

AI referral sessions converted at 7.18% against 6.08% for organic sessions, and AI traffic converted worse than organic on 23 of 63 websites.

Source: The SEO Works, "How Much Traffic Do AI Assistants Actually Send, And Does It Convert Better?", 63 websites meeting minimum traffic thresholds, September 2025 to August 2026

Two further datasets measure other actions. Microsoft Clarity looked at 1,277 publisher and news domains for one month.

LLM referrals showed a 1.34% subscription click-through rate against 0.55% for search, and 1.66% for sign-ups against 0.15%, across 1,277 domains analysed.

Source: Microsoft Clarity, "AI Traffic Converts at 3x the Rate of Other Channels (Study)", Ihab Rizk, one month of data, 6 November 2025

Ahrefs reported its own site over 30 days.

AI search traffic accounts for just 0.5% of Ahrefs' traffic for the last 30 days, and 12.1% of Ahrefs signups come from AI search.

Source: Ahrefs, "Does AI Search Traffic Convert Better Than Traditional Search? For Ahrefs, Yes: 0.5% of Visitors Drove 12.1% of Signups", Patrick Stox, Ahrefs' own site, 30 days, 16 June 2025

The table sets the four side by side.

Study Sample Period Result for AI visitors
Marie Haynes 55 sites tracking conversions 12 months 13.8% against 3.02% for Google organic
The SEO Works 63 sites Sept 2025 to Aug 2026 7.18% against 6.08%, lower on 23 sites
Microsoft Clarity 1,277 publisher and news domains One month Subscription click-through 1.34% against 0.55%
Ahrefs Its own site 30 days 0.5% of traffic, 12.1% of signups

What the numbers leave open

Each study defines a conversion in its own way: a key event in GA4, a subscription click, a signup. Samples come from agencies, publishers and one software company, and several studies are published by firms that sell SEO services or tools. The common finding is a small channel whose visitors often act more than the average visitor, and the size of that effect changes from site to site. A plausible reason is that these visitors compared options in the assistant before clicking, so they arrive with a shortlist already made. The studies do not test that explanation.

The named samples are American and French, Pew's 900 US adults and Indig's US and France domains, and we found no figures for Spain, Belgium or the Netherlands in these studies. A company selling into several language markets should read them as a starting hypothesis and run the test on its own pages.

What to do with this

Three actions follow from the evidence, in this order.

  1. Measure the channel. Group the referrers from chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com in your analytics, and add "an AI assistant" to the contact form question about how people found you. Our conversion tracking service sets this up per market.
  2. Make the cited pages convert. A visitor who arrives from an AI answer has already read a summary of your offer, so the page needs a clear next step, the same facts the answer quoted and a short path to an enquiry.
  3. Test the questions your buyers ask. Run them in each language you sell in and record which brands, sources and pages appear. The search everywhere strategy lists the surfaces worth checking.

Where this leads

See what generative engine optimization looks like on your site

We start with a thirty minute call on your markets, what already ranks and what has already been tried. Questions come before recommendations, and what comes back afterwards is a written scope naming real pages and deliverables.