Playful interactions have the potential to resolve some of the shortcomings inherent in collaboration in VR space, and offer a potential path for improving collaboration by taking greater advantage of the wide range of interactions possible in the virtual environment. Building toward this goal, this thesis experimentally evaluates two potential approaches to developing social VR collaborative experiences. The first approach offers a visual abstraction of conversational state, allowing participants to see the recent history of their conversation and adjust accordingly. The other introduces an ambient agent which does not participate in the conversation or provide any prescriptive advice, but instead passively indicates whether the conversations has become "imbalanced". They were evaluated through user tests involving a collaborative problem solving task in order to establish their effectiveness in social VR situations.