We've rolled out several key improvements to our AI tools. Users can now apply structural filters when searching data, specifying filenames, sheets, and row ranges for more precise results. This enhancement allows for better control over the data used in AI-powered tasks.

To improve the AI Fill experience, we've implemented measures to prevent the AI from repeatedly using the same data chunks. This ensures that subsequent searches return fresh, relevant results. Additionally, we've refined the way get_file_content handles large files, ensuring that the tool remains performant by capping the amount of data processed in a single call. This prevents potential token overages and maintains responsiveness.

We've also enhanced our monitoring and diagnostics capabilities by adding detailed logging for retrieval tools. These logs provide valuable insights into how the AI tools are functioning and help us identify areas for further optimization. These updates contribute to a more robust and efficient AI-powered data workspace.