ChatGPT Referrals Grew 43%. Could They Replace Lost Google Traffic? | 2026 Research + Free Template
ChatGPT is sending more visits to websites. Is the increase large enough to make up for falling Google traffic? Allua recalculated Ahrefs’ fixed panel of 118,671 sites, updated on 11 November 2026. Between October 2025 and October 2026, ChatGPT referrals grew about 43%. The roughly 1.29 million additional visits were equivalent to around 1% of the observed decline in Google-sourced traffic. This compares the size of the changes across channels. It cannot tell us where individual users went. Ahrefs benchmark
1. How wide is the gap between Google and ChatGPT?
Source | Oct 2025 estimate | Oct 2026 | Year on year |
~430.0M | 297.8M | −30.7% | |
ChatGPT | ~3.01M | 4.30M | +43.0% |
We used October 2026 as the anchor and worked backwards using the published month-on-month changes. The earlier volumes and year-on-year changes are estimates because the source rounds its figures. Allowing for rounding in the monthly rates and anchor volumes gives a ratio of roughly 0.93% to 1.02%. In the same October 2026 panel, all search engines represented 27.55% of traffic and identifiable AI assistants represented 0.46%. Search engines sent about 60 times as much traffic. Ahrefs channel and monthly data

Each line starts at 100 in October 2025 to compare growth rates. Use the table to compare absolute traffic. The analysis makes no adjustments for market, industry or seasonality and cannot attribute Google’s decline to AI summaries, changing demand or other changes to search.
2. Do people still click when Google shows an AI summary?
Pew Research Center examined browsing activity from 900 US adults in March 2025, covering 68,879 Google searches, and reran those queries in April to identify summaries. Traditional-result clicks occurred on 8% of visits with a summary and 15% without one, a difference of seven percentage points. Links inside summaries were clicked on about 1% of visits with a summary. The query groups differed, so this is an observed difference. Pew click study
A separate Ahrefs desktop study compared 300,000 keywords in December 2023 and December 2025. After adjusting for the change in the comparison group, the December 2025 position-one CTR for queries with an AI Overview was about 1.57%, around 58% below the estimated baseline of 3.73%. The estimate applies to a specific sample and ranking position. Use it to guide what you check in your own CTR data. Ahrefs CTR study
3. Which pages should you check first?
Search intent | Hong Kong example | Check first |
Informational | what is an AI Overview | CTR at comparable ranks and summary frequency |
Commercial comparison | Hong Kong SEO agency pricing | Comparison-page clicks and qualified enquiries |
Transactional or local | book an SEO audit Hong Kong | Form completion and qualified enquiries |
Navigational | Allua contact | Branded CTR and correct landing page |
In Pew’s US sample, 8% of one- or two-word searches triggered summaries, compared with 53% of searches with ten or more words. Query length can help you choose what to sample, particularly for question-led content. For Hong Kong, review Traditional Chinese and English queries separately and classify intent and AI summary presence from the actual results. Pew query analysis
4. What if AI referrals convert better?
In Adobe’s September 2026 US retail data, AI-referred visits had a 90% higher purchase conversion rate and 83% higher revenue per visit than non-AI traffic. Audience and channel mix can differ between the groups. Contentsquare’s 2026 benchmark recorded 632% year-on-year growth in AI referrals, with a 0.2% share of total traffic in Q4. Hong Kong service businesses should measure both identifiable AI sessions and qualified enquiries. Adobe retail data | Contentsquare channel benchmark
For example, losing 200 organic visits at a 2% qualified-enquiry rate implies four fewer enquiries. Adding 50 AI-referred visits at a 4% rate would recover two. At that rate, you would need about 100 additional AI visits in total to recover four. Replace the assumed rates with your own measured rates.
Free AI search traffic tracking template
Track organic Google clicks and impressions, Google AI feature impressions, AI-assistant referral sessions and qualified enquiries in one place. The Word template includes a worked row and guidance for missing data.
How Hong Kong businesses can measure this
1. Compare two complete 28-day periods in Search Console. Break out market, language, device and landing page. For key pages, record ranking and CTR. If clicks fall while rankings hold, check the actual results, search demand and page layout.
2. If available in your account, use Search Console’s generative AI performance report for AI Overviews and AI Mode impressions. These impressions are already included in overall Web performance. The impressions report does not isolate AI feature clicks. Google Search Console documentation
3. Group identifiable AI-assistant referrals by session source in GA4, such as chatgpt.com, perplexity.ai and claude.ai. Compare them with qualified enquiries in your CRM. Keep missing referrers unattributed and apply the same rules across periods.
To identify the pages most exposed to traffic loss and worth improving first, tell Allua your website, target market and most important type of enquiry.




Comments