While CRTs are the gold standard for establishing causality, that doesn't mean all is lost in observational data alone. Causal inference is a rich field, with a lot of work in recent decades. There's so much more to causal inference than the old dismissive chesnut about correlation.