AI teams often face challenges in selecting the right data for training models, and recent updates aim to enhance data visualization and exploration to improve decision-making in organizing and prioritizing data. Leveraging the Catalog foundation, users can now efficiently search, explore, and browse large-scale public datasets, gaining inspiration for their own AI pipelines. The updates include natural language search capabilities, allowing users to query data rows by metadata, annotations, and other filters, thereby capturing the complexity of specific use cases and browsing vast data stores. Natural language search, powered by CLIP embeddings, enables rapid insights by returning search results for millions of assets in seconds, and it can refine searches with detailed queries such as finding medical devices in x-ray images. Additionally, the "Find text" filter allows users to search for specific phrases across various media types, enhancing the ability to manage and understand diverse datasets efficiently.