Does AI referral traffic convert better? Check where it lands before you trust the ratio

KEY TAKEAWAYS

1. Multi-site studies put AI referrals near 1.2 times organic at the median, with a wide spread between sites.
2. In one dataset the same AI traffic read 0.67 times paid search and 2.48 times direct. The baseline sets the headline.
3. Landing page mix explains much of the raw gap. Compare like pages with like before acting.
4. Most B2B sites have too few AI conversions for a page-level verdict. Pool, add a range, then check the CRM.

Two datasets released or updated in the past three weeks reach the same conclusion by different routes. Shero Commerce's study of 96.4 million sessions, published on 28 September 2026, found AI referrals converting better than Google organic at the median store. The SEO Works, updated on 14 October, found the same across 63 sites, but below organic on B2B product sites. Older work found the opposite, and single-site figures run as high as 23 times. These studies are not measuring one thing badly. They are measuring different things. This piece shows which choices move the ratio and gives a method for reading your own.

What do the studies actually find?

Here are the main multi-site results, plus the one single-site figure that gets quoted most.

Two things hold across the multi-site rows. The median edge, where there is one, is near 1.2 times. And the spread is wide. AI converted worse than organic on 37% of The SEO Works' sites, and Siege found AI matching or beating organic on 72%, which leaves 28% below.

Why does the same traffic look better or worse?

Three choices do most of the work.

The baseline. The SEO Works ran its AI traffic against four channels. It came out at 0.67 times paid search, 1.18 times organic, 1.37 times all sessions and 2.48 times direct. Same visitors, four headlines. Adobe compares with all non-AI traffic, which mixes email, paid search and organic. The organic baseline has its own problem. Google's channel definitions place clicks from AI Overviews and AI Mode inside Organic Search, so part of the comparison group is AI-assisted too. We covered that reporting gap in what an AI impression is worth.

The period. Kaiser and Schulze's data ends in July 2025. Adobe's retail series shows AI traffic converting 38% worse than other traffic in March 2025 and 42% better a year later. Shero's AI conversion rate rose from 0.66% to 1.75% between September 2025 and September 2026. The 2024 to 2025 study is not wrong. It describes an earlier channel.

The landing page. This is the choice most summaries skip, and the one that matters most for a funnel.

How much of the gap is where visitors land?

A lot of it. AI visitors do not land where search visitors land.

Shopify reported that about 55% of AI-referred sessions in Q1 2026 started on a product page, against about 20% for organic search. Shero's table shows why that matters. Product pages took 51.0% of AI visits and converted at 1.77%. Homepages took 22.8% and converted at 0.57%. Blog and guide pages took 5.1% and converted at 0.29%. A channel that sends half its visitors straight to one of the best converting pages will look strong, whatever those visitors intend. 180 Marketing also removed homepage and blog traffic before measuring its 31% edge.

The mix also moves on its own. Around 7 May 2026 ChatGPT began turning the sites and brands it names into links. OpenAI did not announce the change. Botpresso's study of 52 sites found the homepage share of ChatGPT referrals rose from 4.4% to 31.2% between matched six-week windows. Previsible's data, reported in Search Engine Land, found 34.6% of AI visits to SaaS sites landing on internal search pages. A channel rate can rise or fall because the landing mix shifted, with no change in who is clicking. The homepage effect is also why the first-party pages audit starts there.

Mix does not explain everything. Shopify compared AI and organic sessions that both started on a product page and still found AI about 49% higher in Q1 2026. CMSWire's account of Shopify's Q2 data gives a larger figure, about 80%, but against organic sessions overall rather than page-matched ones. That is the mix effect again. In ecommerce, then, part of the edge survives once pages are matched. For B2B we found no public comparison that holds the landing page fixed.

What does the B2B evidence say?

Less, and more cautiously. Siege put B2B SaaS at 1.05 times organic. The SEO Works put B2B product sites at 0.62 times pooled, on 10 sites. Amsive's B2B difference was not significant once small sites were removed.

TripleDart's Q1 2026 study is the clearest look at a B2B funnel. Across 28 SaaS companies it counted 65,583 AI-referred sessions and 82 attributable leads in HubSpot. The homepage drew 31.9% of sessions and 67 of the 82 leads, a 0.32% rate. Generic blog posts drew 12,135 sessions and no leads. Attribution was single-touch, so earlier influence is missed.

Lead quality data is thinner still. Goodie, which sells AEO software, reported AI search leads closing at a 56.3% higher rate across 13 companies from October 2024 to April 2025, on last-click data. Invoca found 49% of ChatGPT-referred calls became qualified leads but did not publish close rates by channel. Neither is enough to set a planning number.

A worked example: the 2.4x that was 1.08x

A hypothetical B2B SaaS company pulls one quarter. The conversion is a demo request.

The raw ratio is 2.75 / 1.155, or 2.38 times. Now ask what AI visitors would convert at if they landed where organic visitors land. Weight each AI rate by organic's share of sessions:

Mix-adjusted AI rate = (0.05 x 5.0%) + (0.15 x 3.0%) + (0.70 x 0.5%) + (0.10 x 2.0%) = 1.25%

Against organic's 1.155%, that is 1.08 times. Most of the apparent edge came from AI visitors landing on the pricing page.

Then check the one row that still differs. A 5% rate on 400 sessions has a 95% range of roughly 2.9% to 7.1%, using p ± 1.96 x √(p(1 - p) / n). Organic's 4.5% on 2,000 sessions runs from about 3.6% to 5.4%. The ranges overlap. On these numbers the company cannot say AI visitors convert better on pricing at all. It can say the pricing page deserves more AI-referred visitors.

What should you do with your own numbers?

A seven-step method

1. Define AI sessions. Use GA4's AI Assistant channel from 13 May 2026. For earlier months, use a source rule. ChatGPT adds utm_source=chatgpt.com to its links, which helps. Visits with no referrer will still sit in Direct.
2. Fix one conversion. Pick the event that matters, such as a demo request. Check its counting method. GA4 offers once per event or once per session, and migrated goals default to once per session.
3. Choose the baseline. Non-branded organic is closest, if you can isolate it. Never use direct or all traffic. Note that organic already contains AI Overview clicks.
4. Group landing pages into four to six types. Pull sessions and conversions for each type, for both channels, over the same dates.
5. Compute the raw ratio and the mix-adjusted ratio. Report both.
6. Put a range on any page type with fewer than about 30 conversions. If the ranges overlap, call it a tie.
7. Follow leads into the CRM each quarter. Compare lead-to-opportunity rate by original source. Fast, complete follow-up matters more than the channel label, as our note on demo request follow-up argues.

Where this stops

Most of the large datasets are ecommerce. Apart from the Marketing Science paper, every multi-site study here comes from an agency or vendor that sells related services, and none is a random sample. Referral data misses AI-influenced visits that arrive by branded search or with no referrer, so every AI count above is a floor. The 30-conversion line is our rule of thumb, not a published threshold. Replace each borrowed rate with your own within two quarters.

If you want the same habit applied to how AI engines describe your brand before anyone clicks, the waitlist is open.

Questions

What is a good conversion rate for AI referral traffic?

There is no fixed number. Read it as a ratio to your own organic rate. Siege Media found a median of 1.26 times organic across 78 sites in early 2026, and The SEO Works found 1.18 pooled across 63 sites. B2B SaaS sat near parity at 1.05, and B2B product sites below it at 0.62 pooled.

Why did one study find ChatGPT traffic converts worse?

Kaiser and Schulze studied 973 ecommerce sites from August 2024 to July 2025 and found ChatGPT ranked above only paid social. The channel has changed since. Adobe's retail data moved from AI converting 38% worse than other traffic in March 2025 to 42% better in March 2026.

Does Google Analytics track AI referral traffic now?

Yes. Since 13 May 2026 GA4 has an AI Assistant default channel for recognised chatbot referrers. Older data is not reclassified. Clicks from Google AI Overviews and AI Mode stay in Organic Search, and visits that arrive without a referrer still land in Direct.

How much AI traffic do I need before the conversion rate means anything?

More than most B2B sites have per page. At a 5% rate on 400 sessions, the 95% range runs from about 2.9% to 7.1%. TripleDart's study of 28 B2B SaaS companies found 82 AI-attributed leads in a quarter, about three per company.

Sources

1. Shero Commerce, What 96 Million Sessions Show About AI Referral Traffic and Conversions sherocommerce.com
2. The SEO Works, AI Referral Traffic 2026, How Much, And Does It Convert? seoworks.co.uk
3. Marketing Science, ChatGPT Referrals to E-Commerce Websites, How Do LLMs Compare Against Traditional Channels? pubsonline.informs.org
4. Digiday, E-commerce sites see low sales from ChatGPT traffic, new study finds digiday.com
5. Siege Media, AI Traffic Conversion Rates, Are They Really 7x Better? siegemedia.com
6. Amsive, Does LLM Traffic Convert Better Than Organic? A New Data-Backed Study amsive.com
7. 180 Marketing, ChatGPT Traffic Converts 31% Better than Non-Branded Organic Search 180marketing.com
8. Adobe, AI traffic grows but retail sites lag in AI search visibility business.adobe.com
9. Adobe Digital Insights, AI Traffic Trends Report, August 2026 business.adobe.com
10. Shopify, AI-referred shoppers convert better and spend more shopify.com
11. CMSWire, Does AI Search Really Beat Organic? Not Yet, Shopify Says cmswire.com
12. Botpresso, 98% Growth in ChatGPT Referral Traffic After the May 2026 Links Update botpresso.com
13. Search Engine Land, ChatGPT commands 92% of AI referral traffic. Here's what 6.77 million sessions reveal searchengineland.com
14. TripleDart, What Page Types Do LLMs Cite? A Study of 28 B2B SaaS Companies tripledart.com
15. Google Analytics Help, Default channel group support.google.com
16. Google Analytics Help, What's new in Google Analytics support.google.com
17. Google Analytics Help, About key event counting method support.google.com
18. Loves Data, AI Assistant Channel in Google Analytics lovesdata.com
19. OpenAI Help Center, Publishers and developers FAQ help.openai.com
20. Search Engine Journal, Google Analytics Adds AI Assistant As Default Channel Group searchenginejournal.com
21. Ahrefs, Does AI Search Traffic Convert Better Than Traditional Search? ahrefs.com
22. Goodie, AI Search vs Traditional Search, Which Drives Better B2B Pipeline? higoodie.com
23. Marketing Dive (Invoca press release), New Invoca Data Finds the Best Leads Now Start in ChatGPT marketingdive.com

NEXT STEP

See where your brand stands.

Your Digital Authority Score, who AI recommends for your target questions, and the placements that close each gap. Your first report is free.

Get your free report →