A goal of behavior-analytic interventions is to produce behavior that is maintained under naturalistic conditions. In this experiment, we studied the effects of a speaker immersion protocol (SIP) on the number of speaker responses (tacts and mands) emitted by 3 preschool students under naturalistic, not directly targeted, conditions. During the SIP, the researchers provided 100 daily opportunities for the participants to emit mands using the target mand form by contriving establishing operations (EOs) throughout the school day. The effects of the intervention were evaluated using a multiple-probe design by measuring target mands during EO probe sessions and the number of mands and tacts emitted during noninstructional-setting probe sessions. The researchers found that the SIP produced increases in both targeted and generalized verbal behavior.