Version 6.1
September 28, 2026
Enhancements
Pass and Fail Scoring for Evaluation Forms
Quality managers can now define a pass score for a form as a whole and for each of its questions, so an evaluation returns a clear pass rather than a score to interpret. A new "Pass / Fail" section in the form builder's left panel carries "Form" and "Question" checkboxes, both selected by default, each with a comparison operator and a required "Pass score". Where question pass scores are set, the Question-Based report now shows how often each question failed, in the form "Failed: 18 of 24", and totals failures on every section heading.
Configurable Conversation Limit in Ask GenAI
Analysts can now set how many conversations an Ask GenAI project analyses, replacing the fixed limit of 500 that applied to every project. A "Max conversations" slider with a typed input sits under "Language" in the project panel and accepts any value from 50 to 1000, with 500 still the starting point for a new project. Below 100 the panel warns that AI patterns may be statistically unreliable at that sample size.
Choosing What Automatic Tagging Reads
Administrators can now decide what an automatic tagging configuration sends to the model, so tagging can work from a conversation's topic and summary rather than the transcript alone. A required "Tagging source" multiselect above "Input prompt" offers Transcript, Topic and Summary, and defaults to Transcript. The field appears only when the tenant has AI Insights and automatic summarization switched on; where it is hidden, automatic tagging continues to run on the transcript.
Vertex AI as a Generative AI Provider
Vertex AI can now be selected when creating an LLM configuration, so organizations standardized on Vertex AI can drive the platform's generative AI features through their own deployment. Choose "Vertex AI" in the Provider list under Settings > Generative AI > Configurations and supply the project ID, location, private key, client email and token URI. "Fetch models" then lists the models available on that project, as it does for the other providers.
Reindexing Categories That Include the One You Edited
When a category that other categories merge in is saved, the platform now offers to reindex those parent categories together instead of triggering each one by hand. A "Reindex related categories" dialog lists every category the saved one feeds into, with an indexing period and the date each was last used, and reindexes only the ones selected. The period cannot start earlier than the first day of the previous month, which keeps the extra batch processing bounded.
Queueing Quick Searches at the Concurrency Limit
Quick search now limits how many searches a single user can have running at the same time, so several long date ranges no longer slow indexing for everyone on the tenant. A user who reaches the limit meets a "Concurrent search limit reached" dialog offering "Add to queue" or "Cancel oldest & start", and a queued search starts on its own as soon as a slot frees up. The limit is a tenant setting an administrator can adjust.
Per-Question Agent Performance Export
Quality managers can now export agent performance broken down by question, which was previously visible only as averages on screen. An "Export results" button on the Question based tab of AQM Reports > Agent produces a workbook with one row per agent, question and evaluation, carrying the agent's score alongside the department and organization averages for that question. The export runs in the background, and a notification links to the file when it is ready.
Attached Data in the Evaluation Results Export
Evaluation results can now be exported with any number of attached data fields, instead of the two fixed slots that forced repeated exports to cover them all. The "Export results" modal replaces "Attached data 1" and "Attached data 2" with a single multiselect that includes a "Select all" option, and each selection becomes its own column after the Comment column. Nothing is selected when the modal opens, so an export made without a choice carries no attached data columns.
Category Groups Sheet in the Interactive Reports Export
The Interactive Reports export now includes a "Category groups" sheet, so category group analysis can be worked on in Excel rather than read on screen. Each group appears with its total count and share of calls, followed by its categories with their own counts, share of total calls and share within the group. The sheet's total conversation count covers grouped conversations only, so it will not match the conversation count sheet in the same file.
Department Hierarchy in Excel Exports
Department cells in Excel exports now carry the full path in "Parent / Child" form, so exported data can be grouped by hierarchy without a conversion step. This applies to the agent performance, interactive reports department distribution and evaluation results exports, and to the conversation list and user list exports as well. A department with no parent still exports as a single name, and an unassigned department leaves the cell empty.
Usernames Beside Agent Names
Agents who share a full name can now be told apart at the point of selection: agent pickers and grids show the username beneath the full name, prefixed with "@" and in a smaller size. Excel exports that carry a full name column now carry the username with it. Search already matched on usernames, so only what is displayed has changed.
Server-to-Server SSO Token Exchange
Integrations that sign users in through an external identity provider can now exchange a validated id_token for a single-use handoff code over a back-channel call, so the token never has to reach browser JavaScript. A new endpoint, POST /api/sso/token-exchange, accepts the id_token and the tenant and returns a short-lived code that can be redeemed once. The existing token endpoint is unchanged, so SSO bridges built on it keep working.
Improvements
Category Group Colours and Names
The colour chosen in the category group form is now used for that group in the category group distribution diagram under Analytics > Reports, including where the default colour has been left in place. Group names must also be unique within the tenant: saving a name that already exists is rejected with "Category group name is already in use." Groups that already share a name are left untouched.
Categories Without "Include in reports" in Filters and Alerts
Every active category can now be selected in the Reports filter panel and in Coaching & Training email alerts, not only those marked "Include in reports". Categories built for other purposes can therefore be reported on and alerted from without changing the category itself. Existing alerts keep the categories they already have.
Licence Recognition After a Subscription Change
A subscription edited in the Portal is now reflected in licensing straight away, instead of leaving a window of up to about fifteen minutes in which the platform could wrongly report that a licence was not found. The licence lookup that follows a tenant update bypasses the licence service cache, and that is recorded in the logs. Every other licence lookup is unchanged and continues to be served from the cache.
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.