Framed Autonomy in AI-Augmented Business Process Management: Vision, Opportunities and Challenges AI-augmented Business Process Management Systems (ABPMSs) are an emerging class of process-aware information systems empowered by AI technology for autonomously unfolding and adapting the execution flow of business processes within a set of potentially conflicting constraints, called process framing. Process framing entails establishing multiple constraints encompassing procedural rules, best practices, policies,and norms that must be considered during the execution of a business process. This enables the key feature of framed autonomy: an ABPMS can autonomously decide how to progress the execution, as long as the boundaries imposed by the frame are respected. While for a conventional BPMS the concept of framed autonomy is rigidly tied to the prescriptive interpretation and enactment of a predefined process model, an ABPMS must simultaneously account for multiple (potentially conflicting) constraints regulating strict and/or flexible aspects of a process execution in a procedural or declarative way. There are, however, several challenges that need to be tackled to properly achieve framed autonomy. In this talk, I concentrate on three particularly relevant challenges: (1) the need of detecting and handling behaviors that break the boundaries, deviating from what is expected; (2) the need of considering uncertain boundaries, where constraints may not necessarily always hold; (3) the need of supporting dynamic conversations between the ABPMS and the process executors to inform them of the process progression, and make recommendations to improve process performance