My site can enhance itself to client side rendered (CSR) (ish) if a prediction can be made by mouse movement analysis. It can gauge a visitor's familiarity with the site, confusion, etc. The content gets adjusted to appeal to the individual without being intrusive.
Problem is, Googlebot just "clicks" links without triggering mouse movement. Without usable movement data, the algorithm can't predict intelligently, so CSR isn't used. The session remains server side rendered (SSR).
The content google sees is valid. New sessions always start out SSR. Mouse movement is tracked until a prediction can be made. From there CSR takes over until they leave the site.
Is this allowed by Google standards? I would think so since the Google cached version and the version served to someone entering my site from Google matches.
Is it cheating to serve different versions of the same content to users and crawlers? Not a duplicate of but the same general issue.