Michael Blume writes concerning the “personification of the universe” model of religion:

Religious traditions seem to derive their motivational, cooperative and then reproductive potentials from the belief in superempirical agents – ranging from deceased ancestors to various spirits, angels and demons to gods, bodhisatvas and alien visitors from outer space to God. . .   In fact, non-personal systems such as early Buddhism, Jainism, Taoism (etc.) had to adopt superempirical agents (such as bodhisatvas, khami, tirthankaras, the Lord Tao and many more) in order to survive demographically. The underlying logic is rooted into evolutionary theory itself: As human beings, we might be ready to accept commandments from supreme “personalities”- but not from abstract and non-living objects or principles.  As Friedrich August von Hayek rightfully observed: A theistic commandment such as “Be fruiftul and multiply” (Genesis 1, 28) may be accepted by religious believers as authoritative and even beneficial, although it cannot be verified empirically. By personification, religion is able to attribute value to forming families and having children.  In contrast, to accept empirically tested hypotheses as “teaching” normative commandments would constitute a natural fallacy contradicting our evolved feelings as well as philosophical lore. Although modern definitions of Darwinian or Evolutionary Fitness agree on the importance of reproductive success in evolutionary processes, we are simply not ready to accept any “commandments” thereof.

I’m pretty well convinced myself that when it comes to religion, agency is where it’s at– even when official theology says otherwise. If you want to read a good defense of that point, I would highly recommend Jason Slone’s book Theological Correctness: Why Religious People Believe What They Shouldn’t.  He marshals a great deal of evidence from experimental psychology to show that people who are thinking religiously are thinking in terms of agency, and that they think of supernatural agents (gods, spirits, bodhisattvas, and so on) as essentially being like humans with some superpowers tacked on.

The argument is basically that religion is ultimately about intuitions– that theology is something believers accept to signal that they are part of their own particular religious community, but their reflexive reactions suggest that there are more deeply-rooted, fundamental ways of thinking about agency which are applied when thinking about the supernatural.  In other words, that given the right circumstances you can “tease out” convictions in people about how supernatural agents think and behave that rely more on ordinary intuitions about how humans think and behave than on the particular dogma that defines one’s faith.