AI Search Traffic Studies: What Three 2026 Analyses Found

Three Studies Asked Different Questions

Two researchers counted visits referred by AI assistants. The third investigated whether AI exposure might lead to Direct visits or branded search activity. Their results belong alongside one another, not in a single pooled traffic estimate.

Haynes compared AI referrals with Google Organic, while The SEO Works used its measured organic search sessions as the denominator. The 0.73% and 2.6% figures also come from different sites, periods and referral definitions. Neither study isolates which of those differences explains the gap. These are sample-specific ratios, not estimates of AI search’s share of all discovery activity.

Indig’s result is even less interchangeable with those percentages. His investigation included 632 correlation tests across four brands and found that promising relationships weakened under alternative time alignments. It does not show that indirect AI influence is absent. It shows how difficult that influence was to establish in the data examined.

Higher Conversion Rates, Smaller Totals

In Haynes’s 55 properties with conversion tracking, ChatGPT visitors converted at 13.8%, compared with 3.02% for Google organic visitors. Yet Google organic generated 27 times more conversions than all identified LLM referrals combined in that dataset. A higher rate did not overcome the difference in volume.

The SEO Works reported a smaller conversion-rate advantage: 7.18% for AI assistant referrals versus 6.08% for organic search. Across its reported comparison, AI referrals generated 5,001 conversions, against 155,605 from organic. The agency restricted its conversion analysis to 63 qualifying websites, rather than all 95 in the traffic sample.

The average conceals an important exception. On 23 of those 63 sites, AI referrals converted worse than organic visits. For the ten B2B product websites, the pooled AI-to-organic conversion-rate ratio was 0.62, despite those sites receiving a relatively high share of AI referral sessions. That is a stronger reason to check your own industry and conversion definition than to assume AI visitors always arrive with better buying intent.

These figures also depend on what each GA4 property records as a conversion or key event. The SEO Works notes that events are not necessarily unique customers or completed sales. A referral study can establish a measured channel comparison without demonstrating incremental revenue from a specific optimisation programme.

What the Results Mean for SEO Decisions

First, do not treat traffic from third-party assistants as the whole of AI search. Google’s documentation on AI features says clicks from AI Overviews and AI Mode are included in Search Console’s Web search performance data. Visits from ChatGPT, Perplexity and similar services may instead appear as identifiable referrals in GA4. These reporting views answer different questions and cannot simply be added together as though they were mutually exclusive measures of AI influence.

Second, separate a measurable visit from a possible contribution to discovery. Indig found that unusual Direct traffic increases could be dominated by automated activity. He also could not confirm that changes in AI mentions were consistently followed by higher branded search or Direct traffic. Neither result demonstrates that an AI mention caused, or failed to cause, an eventual sale.

For an SEO team deciding where to spend next, the order of work should reflect that uncertainty:

• Protect organic search work that already delivers qualified visits or revenue. Small AI referral totals alone are not a reason to replace that foundation. Google also says established SEO best practices remain relevant for AI Overviews and AI Mode.
• Segment AI assistant referrals by source, landing page, market and conversion type. Check engagement and suspicious Direct traffic before treating a spike as demand.
• When testing AI visibility work, document the cost, a consistent set of prompts or exposure measures, customer-reported discovery and qualified business outcomes. Evaluate them against a baseline before increasing the budget.

This is a measurement decision, not a universal verdict against GEO. For a fuller treatment of how to connect AI referrals with lead quality and attribution, see MOCOBIN’s AI search attribution analysis. The SEO reporting guide covers the underlying distinction between Search Console search metrics and GA4 website sessions.

The next useful check is not whether AI traffic exceeds a published benchmark. It is whether identified AI visits, assisted discovery evidence and qualified outcomes justify the next unit of investment for your own site.