Long-term therapeutic processes often involve large amounts of fragmented information distributed across months or even years of sessions. Therapists frequently work with handwritten notes, session summaries, and recurring patient references that can become difficult to organize, retrieve, and connect over time.

In this workshop, we explore how AI systems can support therapists by improving the handling of long-term therapeutic information. We discuss approaches for transforming handwritten or spoken session notes into usable text, retrieving relevant past information across therapy histories, and generating contextual summaries to support therapeutic workflows. The workshop also presents initial directions for integrating long-context retrieval and memory systems into clinical documentation pipelines.

Rather than focusing on replacing therapists, the goal is to investigate how AI can support existing therapeutic practices through better information organization, retrieval, and contextual assistance. We will also present early feedback collected from therapists and discuss practical challenges, opportunities, and future directions for AI-assisted therapeutic workflows.