V2.0 Release Notes: Toro AI Assistant

We are excited to announce a significant update to the Toro AI Assistant. This release focuses on increasing search precision, giving you more control over the chat experience, and deeper integration with your Research Management System (RMS) workflow.

Smarter Search & Filtering

We have upgraded Toro’s ability to understand and process complex data queries, ensuring you get exactly the results you are looking for.

  • Improved Logic (AND/OR): Toro now intelligently applies "AND" vs "OR" logic based on natural language context. This results in much more intuitive search results when asking for documents that meet multiple criteria simultaneously.
    Example:
    What are the latest notes about Fantastic Fund and Salesforce?” – in this example Toro would search for both Fantastic Fund AND Salesforce but would recognize that there are a few different spellings for Salesforce (e.g. Salesforce, Salesforce Inc., etc.) so it would perform an OR operation for those differences.
  • Full List Field Options Visibility: Previously, Toro could only "see" a max of 10 items in a dropdown list. It now has access to all values in your list fields, ensuring no category or tag is overlooked during a search.
  • "Field Presence" Filtering: You can now ask Toro to find documents specifically based on whether a field exists. The search planner now recognizes when a query implies a check for a field's presence (because we plan to extract the values). In such cases, it generates a filter that returns all documents containing the specified field, regardless of its value. This enables a later step in the process to reliably extract the field's content for the user.
    Examples:
    Show me all notes with an attendee“– in this example, Toro would search for fields that could have an attendee associated.
    What are the values from PMS and list a summary of the related document“ – in this example we have a “list field” called PMS where we want to know all the potential values of that list field along with any Research items that are linked to each list field item.
  • Increased Accuracy: We have refined the suggestion engine to filter out "hallucinated" or non-existent field values. Toro will now only suggest and query against field values that actually exist in your database.

Enhanced Usability & Control

We have made interacting with Toro smoother and more flexible to suit your multitasking needs.

  • Background Processing (Minimize Window): You can now minimize the Toro window while it is generating a response without disconnecting the stream. Feel free to continue working elsewhere in the application while Toro prepares your answer in the background. Colla