I think Prolog suffers in that comparison mostly because of its much more ambitious scope. Most non-developer/DBA people have no concept of what a SQL query is

I think Prolog suffers in that comparison mostly because of its much more ambitious scope. Most non-developer/DBA people have no concept of what a SQL query is actually doing, whereas most nontrivial Prolog programs require conceptualizing the depth-first-search you're asking the language to perform in order to get it right. If you restricted your Prolog world to the kind of "do some inference on a simple family tree database of facts" that people first learn, Prolog would be pretty easy too.