What models of causal inference do you have in mind that is not amenable to DAGs?

The reason DAGs are useful because it shares similar axioms as probability.