Agentic AI for Music Knowledge Representation: Towards the Automated Generation of IEEE 1599 Documents

IEEE 1599 is a multilayer format designed to represent and synchronize heterogeneous music information, including symbolic notation, audio recordings, metadata, and analytical descriptions. Despite its expressive power, the creation of IEEE 1599 documents remains a labor-intensive process that often requires the integration of data originating from multiple sources and processing pipelines. Recent advances in Artificial Intelligence (AI), particularly in the field of AI-based agents, offer new opportunities for automating and coordinating such tasks. This contribution discusses the potential role of specialized AI agents in the generation of IEEE 1599 documents. We identify three application scenarios that are particularly relevant to the automatic population of the Logic layer and the synchronization structures of the format. First, conversion agents may transform existing symbolic encodings, such as MusicXML and MEI (Music Encoding Initiative), into IEEE 1599 representations. Second, Optical Music Recognition (OMR) agents may extract symbolic information directly from digitized scores and generate structured music descriptions. Third, audio synchronization agents may automatically align symbolic and audio content, producing synchronization data required by the format. We argue that an agent-based architecture can orchestrate these heterogeneous processes through autonomous task execution, verification, and refinement, thus reducing human intervention while preserving data quality. The paper outlines a conceptual framework for such an architecture and discusses research challenges, including reliability, interoperability, provenance, and validation of generated content. We conclude by highlighting how agentic AI may contribute to the broader vision of intelligent music archives and digital cultural heritage systems.