The point is that we can verify whether something works (with a randomized controlled trial) independently of knowing why it works, and psychotherapy interventions are commonly evaluated with RCTs.