Intelligent Insights Deploy Notes (2025-March)

[#1]

Release notes outline upcoming customer-impacting changes that are expected to be displayed in beta and production environments. Please note that features considered for future development are not included in the notes.

Unless otherwise stated, all features in this release are available in the Production environment as of 2025-March.

• Updated Features
• Ask Your Data
• Large Language Model

Updated Features

Ask Your Data

Large Language Model

Feature Availability	Production environment only
Beta Environment Availability	Not available in the beta environment
Location to Enable Feature	NA
Name of LTI Tool	None
Permissions	Intelligent Insights/Ask Your Data
Related APIs	None
Affects User Interface	Yes
Affected Areas	Large Language Model
Related Ideas	None

Summary

The following updates are made to the Large Language Model:

• Updated model context with questions related to grades posted for assignments, posting policies, announcements, and calendar events and appointment groups.
• Minor improvement to consistency for certain types of questions such as getting user emails, announcements.
• Improved context for schema linking step that helps in cases where the system fails to identify tables needed to answer these questions.
• Improves chat response accuracy by displaying the correct number of rows in the Summary section for suggested questions.
• Enhances SQL generation by improving reasoning time before response, resulting in more accurate and higher-quality answers.
• Chat explanations load faster for a smoother experience.
• Updated model context to use current_login_at instead of last_login_at for the most recent login.
• Updated model context to allow for NULL start_at and end_at columns for enrollment terms.
• Updated model context with calculations related to assignment completion rates, quiz submission rates, student performance, and missing assignments.
• Expanded the training data set with questions to: 
• Calculate the percentage of submissions for assignments that were marked late or missing for each term.
• Calculate the quiz submission rate for each course.
• Identify students with missing submissions for any of their assignments.
• Determine the courses with the highest rate of assignments turned in late.
• Compare the performance of students who submit quizzes late versus on time.
• Identify students with missing assignments and a grade below 70 in courses that started at least 3 weeks ago.
• Calculate the completion rate for each assignment and the total number of students who have completed each assignment.