Version 5.36
August 28, 2026
Enhancements
Multiple Tags from One Automatic Tagging Group
A conversation that matches more than one response in an Automatic Tagging group can now receive the tag for every response it matches, so nothing is missed when a single conversation covers several subjects at once. A new "Maximum answer count" field on the group sets how many of its responses may be matched to one conversation, from 1 to 5. It is 1 by default and existing groups keep that value, so tagging behaviour does not change on upgrade until the field is raised. Every assigned tag is listed individually on the conversation detail, the tag filter returns the conversation when any one of them is selected, and the categories and queries that use each new tag are re-evaluated as before.
Period Change Indicators on Trends
Each category selected on the Analytics > Trends chart now carries a badge showing how much it moved over the selected period, so a supervisor can see which categories are improving or deteriorating without comparing the two ends of a line. The badge appears automatically, as a percentage with an arrow, in green for an increase and red for a decrease, and is calculated from the first and last periods with actual data in the range. Hovering it shows the opening and closing values with their dates and the calculation behind them. No badge is shown where the first period is zero, where the range holds only one period, or on the "A Day Before" comparison chart. A category's line can also be hidden by clicking it on the chart, and restored by clicking it again.
Separate Model Configuration for Ask GenAI Steps
Ask GenAI can now run its two steps on different models. Under Administration > Settings > Generative AI > Ask GenAI the existing selector is relabelled "LLM Configuration for Initial Analysis", and a new optional "LLM Configuration for Key Insights" selector has been added; both choose from the configurations already defined and are saved together. Because the initial analysis runs again for every conversation matching the query, it can be pointed at a fast model with no reasoning while the key insights step uses one with the reasoning effort the final synthesis needs, reducing cost and latency. Leaving the key insights field empty keeps the current behaviour, with both steps using the initial analysis configuration, and other GenAI features are unaffected.
Choosing Conversation Types on a Category
Conversation types on the category create and edit modal are now selected with checkboxes instead of a single choice, so a quality manager can build a category for exactly the types they work with. The list of available category items follows that selection and offers only the items the selected conversation types have in common, making it clear up front which items a type such as Email, Video Call or Meeting actually supports. Categories that previously used "All" are migrated: the conversation types in use in the tenant come across ticked, so existing categories keep behaving as before.
Improvements
CSAT Ordering on the Conversation Duration Widget
The ordering toggle on the CSAT widget on the Reports screen now sorts the way it is described. Switching the bars from duration order to descending average CSAT put the lowest averages first; the duration ranges are now listed from the highest average CSAT downwards, so the ranges where satisfaction is strongest appear first. Duration order is unchanged.
Minor UI Enhancements
Several minor improvements have been implemented to enhance the user interface. These enhancements make the interface more intuitive and user-friendly, contributing to a better overall user experience.
Minor Bug Fixes
Various minor bugs have been addressed to improve overall system stability and performance. These fixes result in a smoother and more reliable user experience.