02 / IT · Analysis
Meta Muse request share does not measure web market share
One security firm's observations show AI agent activity, but do not establish user numbers or purchase outcomes.

Meta's Muse sent most of the AI agent requests observed by one security firm, Search Engine Journal reported. For website operators, the findings raise a question about how to interpret automated visits.
According to the report, two companies' AI agents accounted for most requests observed over one month. That concentration describes requests within the firm's observations. It does not establish the size of the agents' user bases.
The report also places most requests on product and search pages rather than checkout pages. The destinations matter alongside the agents sending them: they distinguish product exploration from evidence of purchases.
The scope behind the request share
Search Engine Journal bases Muse's share on requests seen by one security firm. Treating that share as a measure of activity across the entire web extends the finding beyond its stated scope.
The report counts AI agent requests. Reading their concentration as a concentration of users treats two different measures as interchangeable. The observed request share does not establish user numbers.
Product exploration and purchase outcomes
Search Engine Journal reports that most requests went to product and search pages. On that basis, the activity highlighted in the report is closer to product exploration than checkout.
The destination breakdown describes page requests. Using it as evidence of purchase outcomes extends the argument from where requests arrive to whether transactions follow. The distinction remains central to interpreting the findings: request activity shows where agents go, while purchase outcomes require evidence of transactions.