The present research uses a technique that permits unique estimation of both automatic and effortful processes in the question-behavior link. Results show that individuals asked to report behavioral intent (vs. those not asked) are more likely to choose options that are highly accessible and positively valenced, regardless of cognitive resources available at the time of processing. This suggests that the effect of intent questions on subsequent behavior is primarily the result of automatic, as opposed to effortful, processing. Practically, this suggests that efforts to debias this robust effect need to affect nonconscious processes and adjust for the automatic impact of being asked an intention question on respondents' behavior.