The integration of artificial intelligence into augmentative and alternative communication (AAC) systems has the potential to significantly enhance their functionality, but evaluating these AI-powered interfaces poses a substantial challenge. The complexity of human needs and preferences, which are inherently multifaceted and nuanced, cannot be fully captured by current evaluation metrics. This limitation is particularly significant because individuals who rely on AAC systems have diverse and intersectional identities, making it essential to develop more comprehensive assessment tools. Researchers are grappling with the intricacies of designing and evaluating AI-powered AAC interfaces that can effectively meet the varied needs of their users1. The development of more sophisticated evaluation metrics is crucial to ensure that these systems are truly effective and supportive. This matters to practitioners because creating AAC systems that are tailored to individual users' complex needs can greatly impact their quality of life and ability to communicate effectively.