Version 6.2
October 13, 2026
Enhancements
Comparing Several Forms in Agent and Evaluator Performance
Quality managers can now compare agent and evaluator performance across several evaluation forms on one screen, instead of switching the Form filter one form at a time. The Form filter on the Performance tab of AQM Reports > Agent and AQM Reports > Evaluator accepts up to 5 forms, each with its own version, and the selected forms appear as chips under "Showing data for:". With two or more forms selected, scores are weighted averages across them, clicking a row opens its per-form breakdown, and "Export results" adds an "Aggregate (All Forms)" sheet before one sheet per form.
Per-Question Evaluator Performance Export
Following the per-question agent export in 6.1, evaluator performance can now be exported broken down by question. An "Export results" button on the Question based tab of AQM Reports > Evaluator produces one row per evaluator, with form score, evaluation count, calibration count and calibration rate, followed by three columns per question: the evaluator's average, the question's maximum possible score and the organization average. Automatic evaluations share a single "Automatic" row, and a notification links the file when it is ready.
Choosing Category Columns in the Conversation List Export
The conversation list export can now carry only the categories that matter for a report, instead of a column for every active category. Ticking "User defined categories" opens a searchable list with "Select all", and only the ticked categories become columns; with none ticked, every active category is included as before. The choice is remembered for each user and filter set, so an export opened again from the same alert link starts with it.
Category Groups and Tags in Report Filters
Reports can now be narrowed the same way as the Conversations screen. The Category filter offers "Group by" on Reports General, Agent performance, Category and Interactive reports, and on Root cause Statistical analysis and Statistical comparison; it defaults to "None", and the chosen group is remembered on each screen. Reports > Agent performance also gains a "Tag" filter, which includes conversations carrying any of the selected tags.
Form Versions, Reasons and Category Names in the External API
Integrations that build their own reports from evaluation results can now resolve forms, reasons and user-defined categories through the External API alone. GET ext-api/v2/evaluation-results now returns formId and formVersionId, and GET ext-api/v1/forms and v2/forms list every published version of each form under formVersions, newest first. A new GET ext-api/v2/reasons returns each reason's id and text, and GET ext-api/v{1|2}/queries/by-version-id/{queryVersionId} returns a user-defined category's name, even when it is inactive or deleted. Answers are still returned as ids.
Improvements
Saving a Full Vertex AI Private Key
The "Vertex AI Private Key" field under Settings > Generative AI > Configurations now accepts up to 4000 characters, so a complete Vertex AI key can be saved. Previously a 200-character limit prevented the configuration from being saved. Fields for other providers are unchanged.
Deleting a Single Conversation Without Waiting
Deleting a single conversation no longer holds the request open or ends in a timeout. The deletion is queued in the background, as bulk delete already was, and "Delete request has been created" appears straight away. Until the background job has run, the conversation can still appear when a list is refreshed, and no notification is sent when the deletion completes.
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.