Release Notes - 2025.09.03

Welcome to the latest One Model product release update. This article provides an overview of the  product innovations and improvements to be delivered on 3 September 2025.

One AI

• Mode selection is now available in One AI Assistant! You now have the ability to explicitly select a mode–Visualizations, Storyboards, and Answers–instead of relying on AI to determine which to use. AI Mode remains the default, but if you are not getting the desired result, you can take control directly. 
  
  We also plan to expand mode selection for upcoming capabilities, such as asking questions about a chart for deeper analysis and asking general questions.
  
  Available modes are as follows (ref 4026):
  • Use AI - AI determines which of the following modes to use based on your prompt
  • Visualizations - generates a query and chart based on information you provide
  • Storyboards -  retrieves the most relevant Storyboards matching the topics or criteria you specify
  • Answers - provides curated answers to key questions and topics

• In this release, we’ve significantly improved styling and branding support for One AI Assistant. Previously, styling was split between hard-coded settings and branding controls. We’ve now added more options for the Assistant to the site branding page and unified them with sensible defaults. (ref 3933)

• Example prompts in One AI Assistant can now be re-ordered by admins. On the One AI Configuration screen, simply drag and drop example prompts to re-order them and click the Save button to retain the changes. (ref 4004)  

User Experience Improvements

• We've improved the accuracy of the axis scale break indicator on Waterfall Charts. The jagged line now renders at the 'TO' value of the defined break range, providing a more precise visual representation of the compressed axis segment.

Data Integration Improvements

We've made several updates to improve your experience with One Model applications, focusing on the API Runs pages. These changes make the platform smoother, faster, and more reliable.

• API Runs Page
  • The main API Runs page now loads runs as you scroll, so you don't have to wait for the entire list to load at once. (ref 22560)
  • The API Runs page is now much faster. You can now use search and paging to easily find specific tasks within a run. The search function works for task names, status, messages, and tables. We've also made the automatic refresh smarter, so it only updates when necessary. (ref 24905)
  • On the API Run page, failed, ignored, and cancelled tasks are now sorted to appear at the top of the list. This makes it easier to quickly identify and address any issues. (ref 25317)
  • Data source names now load faster in dropdown menus on the API Runs page, speeding up the process of selecting data sources. (ref 25183)
  • The task table on the API Runs details page now has a maximum height, preventing it from taking up too much space and making the page easier to navigate. (ref 14306)
• Data Source Fixes
  • We fixed a bug where Greenhouse data sources were using the computer's local timezone instead of the correct universal time (UTC). This correction ensures that you get the correct and complete information, preventing data from being missed due to time differences and keeping your data pipelines stable. (ref 25259)
     

Minor Improvements & Bugs Squashed

• We resolved an issue with the names of pivoted columns not being recognized in the AI explanation contained in One AI Assistant Answers. (ref 4024)
• We addressed a bug in Machine Learning where the EDA and Results Summary reports could get swapped if multiple runs completed within the same minute. (ref 4025)