We demonstrate the production of spoken output with contextually appropriate intonation in the information-state based dialogue system GoDiS. We exploit the context representation in the information state to determine the information structure of system utterances, which we use to control the intonation of synthesized spoken output.