Google Analytics 4 (GA4) is undercounting traffic from artificial intelligence assistants due to fragmentation across its reporting channels, necessitating manual configuration for accurate measurement, according to industry analysis.
The platform’s native AI Assistant channel, introduced around May 13, 2026, automatically tags sessions from recognized AI assistants, but its reliance on both source and medium leads to incomplete data, experts reported.
This fragmentation means that traffic from a single AI source, such as chatgpt.com, can be split across the AI Assistant, Referral, and Unassigned channels, MarTech reported. Google’s list of recognized AI platforms for the native channel is also dynamic and incomplete, omitting significant sources like Perplexity.
The issue stems from how some AI traffic, particularly from in-app browsers, strips the referrer information. When this occurs, GA4 loses the medium data and categorizes the traffic as ‘(not set)’ or ‘referral’, further obscuring the true volume of AI-driven visits.
To address this, analysts recommend creating a custom channel group within GA4. This custom group can consolidate fragmented AI traffic by matching on source using regular expressions and ignoring the medium, allowing for retroactive data correction and the inclusion of unrecognized AI platforms.
Accurate measurement of AI traffic is important for businesses. Similarweb data indicated that referrals from platforms like ChatGPT convert at a high rate, approximately 7 percent, despite currently representing a small fraction of overall web traffic.
By implementing a custom channel group, organizations can gain a clearer understanding of user engagement originating from AI assistants, which is essential for optimizing digital strategies.
Source: Search Engine Journal
