Next-Gen Table Extraction from Visual Documents: Leveraging Multimodal AI