Abstract: A system and method for managing unstructured data stored across multiple data repositories is disclosed. The method includes accessing unstructured data items from distinct data repositories and generating classification metadata for each unstructured data item using a machine learning-based classifier. A residency pointer is generated for each unstructured data item based on residency metadata that identifies its storage location within the data repositories. An unstructured data metafile is automatically constructed for each unstructured data item. Each unstructured data metafile is assigned to one or more virtual queues based on its classification metadata, enabling organized data management. The method further integrates the virtual queues with a web-based data handling interface that facilitates batch operations on unstructured data items stored across the distinct repositories.