Assistance systems equipped with artificial intelligence are present in many Internet of Things applications to address information overload and provide automation. Our client, a multinational technology giant, was interested in studying how far the usage of assistance systems leads to privacy concerns. Using Choice Based Conjoint (CBC) Analysis, we explored privacy preferences for assistant systems on the Internet of Things and quantified the willingness to pay for various privacy functions. CBC is especially suitable for eliciting user preferences during the early stages of product development, providing our client with valuable insights before building the first prototypes.