Speakers use prosody to resolve ambiguity, but what if prosody cannot make distinctions? We explore (1) how speakers employ prosodic and gestural cues to deal with sentences with ambiguous meanings and (2) what insights the audiovisual resolution of ambiguities offers regarding communicative efficiency and effort. Thirty-two native Chinese speakers were asked to articulate twenty-two ambiguous Mandarin sentences. Half could be semantically differentiated using prosody, and half could not. Firstly, participants articulated all ambiguous sentences spontaneously and provided explanations to a confederate, revealing their dominant interpretations. Secondly, participants articulated the same ambiguous sentences twice, each time guided by a hint suggesting a different meaning. Participants’ prosodic cues and gestures were coded and analyzed. The results showed that for ambiguous sentences that can be prosodically distinguished, participants employed various prosodic cues such as pausing, tones, stress, and speaking rates. Additionally, 51.85% of sentences were accompanied by referential (iconic; pointing) gestures, while 17.33% of sentences were accompanied by non-referential (beat; interactional) gestures. However, when prosodic cues were unable to mark ambiguity, participants resorted to more referential gestures (97.30%) but fewer non-referential gestures (1.28%). In conclusion, speakers adopt a multimodal approach to enhance communicative efficiency while there is a trade-off between modalities.