This study proposes a virtual reality platform for language learners to practice a Target Language (TL) and develop cross-cultural competencies through interaction with peers or an AI-agent (limited scope), followed by a scheme for engaging learners to envision their conversations by disclosing their thoughts, reasoning, feelings, and expectations. This platform, Virtual Reality Conversation Envisioning (VRCE), enables the learners to fully share a contextual, immersive environment, simulated in Virtual Reality (VR), to have free conversations by performing role-plays on proposed topics. VRCE is designed to provide first person views during conversation and third person views during envisioning so that learners can take the role of participants and meta-participants at each phase. Learners envision the conversation individually at certain points during the role-play review. Their envisioning is then shared with their conversational partner to detect the misunderstandings, observe the situation from each other's perspective, and to learn about sociocultural cues that led to the different viewpoints. Participants' envisioning revealed interesting differences in their understanding of a shared situation. Findings showed that VRCE is an effective medium to raise learner collaboration and develop cross-cultural competencies. [For the complete volume of short papers, see ED590612.]