AI & ML Research 24 Hours

Last updated: June 21, 2026, 5:30 PM ET

AI & ML RESEARCH

Understanding how large language models interact with their environment is central to developing more capable AI agents. Analysts are exploring tool calling, detailing the mechanisms by which these models decide to access external data or execute actions. This is critical for enabling AI to move beyond text generation and engage with the real world.

In document processing, researchers are addressing challenges with unstructured data. One approach involves reconstructing PDF table of contents for enterprise document intelligence, particularly when original metadata is missing. Separately, for self-service business intelligence environments, new methods for building date tables are emerging, offering alternatives to traditional DAX coding and streamlining data preparation for analysis.