Fully asynchronous courses optimize for flexibility. Learners can start, pause, and finish on their own schedule. But asynchronous courses feel isolating. Without shared timelines, instructors, or cohorts, most self-paced courses feel impersonal and completion rates suffer. Discussion boards decay, peer interaction disappears, and learning becomes solitary. But asynchronous learning does not have to mean solitary learning. I designed interaction models that sustained meaningful dialogue and felt human without relying on instructors or shared student presence. This approach led to: - **Sustained discussion activity** without instructor moderation - **Learner dialogue through AI-mediated interaction**, rather than one-off posts - **Higher perceived connection**, despite the absence of cohorts or live sessions These outcomes made interaction viable in courses where learners never overlap in time. ### Constraints In synchronous courses, interaction is structural: instructors guide discussion, peers respond in real time. In asynchronous, self-paced courses: - There is no instructor presence. - Learners rarely overlap in time. - Traditional discussion forums collapse. I designed two complementary interaction models to enable dialogue without shared time, instructors, or cohorts. ### Design Decisions #### 1. Reframing discussions as artifacts for future learners Instead of positioning discussions as conversations, I reframed them as _contributions passed forward_. Before launch, I seeded discussion boards with exemplar comments. Learners were then asked to: - Respond to an existing contribution, and - Leave a new prompt or insight for future participants. This shifted discussions from time-bound exchanges to a cumulative knowledge layer. Each learner interacted indirectly with both past and future peers, creating continuity without co-presence. #### 2. Introducing AI-mediated dialogue To address the absence of guided interaction, I piloted AI chatbots trained on course content. The bots functioned as structured discussion partners. Learners engaged in back-and-forth dialogue, responded to prompts, and received feedback. These interactions were graded by the AI, providing accountability and scaffolding without instructor involvement. In asynchronous formats, interaction benefits from being redesigned rather than abandoned. Discussions don’t need instructors or overlapping cohorts; they can be built from artifacts and dialogue accumulated across successive learners. The same approach can be applied in contexts where interaction is typically absent or deprioritized—self-paced programs, rolling enrollments, and large-scale online courses.