The method of reflective equilibrium (RE) is primarily taken to be a method of justification for our beliefs, especially our normative beliefs. In virtue of bringing considered judgments into coherence with normative principles, RE claims those judgments and principles thereby both gain some positive epistemic status. Though the justificatory reading of RE is by far the most common, it is not the only possible way to use the method. Revisiting an influential early paper on RE (Daniels 1979) reveals another possible use: RE as a method of _inquiry_. I argue that RE represents a plausible norm for intellectual inquiry, especially in normative domains, over and above its role in justifying particular beliefs. In particular, I argue that RE is a deeply important norm for non-ideal inquirers attempting to learn normative truths when they do not have access to true normative first principles. To demonstrate this claim, I consider a case study in detail: debates about the proper role of AI ethics principles in technology ethics (e.g. Hagendorff 2020; Whittlestone et al. 2019). The resulting argument is in direct contrast to a standard objection to RE, which claims that RE is far too intellectually demanding for non-ideal agents to use. In contrast, I show in this paper how RE, in conjunction with reasonable norms for evidence gathering, represents one of the few paths forward for moderately rational (but uninformed) non-ideal agents to learn normative truths in technology ethics (and elsewhere).