Version 5.35
August 18, 2026
Automatic language detection is not active by default. The language detection service must be configured, and the feature enabled for each tenant that will use it. Until then, chat, email and mixed-model call conversations that arrive without language information continue to be processed with no language set.
For call conversations, detection also depends on the "When the conversations don't have the language information, use" setting under Settings > Other > Global having no system language selected.
Enhancements
Question-Based Objections in Evaluations
Agents can now object to individual questions in an evaluation instead of only to the form as a whole. The "Objection form" modal offers a choice between "Form-based" and "Question-based"; selecting Question-based lists every question in the form, and ticking a question opens its own comment field, so each disputed question is submitted as a separate objection with its own justification. At least one question and a reason for it are required. When an evaluator accepts a question-based objection, only the disputed questions return for re-evaluation, with their reasons shown in order. Available for manual, automatic and user-based evaluations.
Objection Analysis per Question
A new "Objection Analysis per Question" widget has been added to the "Question based" tab under Reports > Evaluator. For each question it shows the total number of objections, how many were accepted, how many were rejected, and the resulting acceptance rate, sorted by objection count from highest to lowest; questions with no objections are not listed. This lets quality managers see which questions cause the most disagreement between evaluators and agents, and revise the form's criteria on evidence rather than impression.
Excluding Users from Licensing
Administrators can now permanently exclude selected users from licensing under Administration > License Management > User Licenses. Unlike "Deactivate", which only releases the current licence and leaves the user eligible for automatic re-licensing, an excluded user consumes no credit and stays out of licensing until explicitly returned with "Include License"; the "Exclusion Type" column shows which users are excluded. New conversations belonging to an excluded user are refused on arrival even when credit is available, and the rejection is reported distinctly from an exhausted quota so the reason is unambiguous.
CSAT by Conversation Duration
The CSAT widget on the Reports screen now also shows how satisfaction relates to conversation length. Conversation counts are drawn as a grouped bar chart split by the duration ranges defined under Settings > Analytics > Definitions, together with the existing CSAT line. Hovering a bar shows the conversation count, CSAT, Average CSAT, Estimated CSAT and Average estimated CSAT for that period and duration range only. A toggle beside the widget title switches the bars between duration order and descending average CSAT. The widget follows the Reports filter panel and picks up changes to the duration ranges on its next load.
Automatic Language Detection for Chat, Email and Mixed-Model Calls
Conversations that arrive without language information can now have their language determined automatically. Where language detection is enabled for the tenant, chat and email conversations with no language, and calls processed with mixed models, go through a detection step that samples the conversation's longest turns and records the most frequent result as its dominant language. Tenants without the feature keep the current behaviour and the language remains empty. If detection is unavailable, the conversation is processed as before with no language set, so analysis is never held up.
User Group Filter on Reporting and Evaluation Screens
The User group filter is now available on the screens that previously offered only a Department filter, including the Reports tabs, Trends, the Agent report tabs, user-based evaluation and Evaluation Results. It works the same way it does on the Conversations screen, so a supervisor can isolate one group's conversations or evaluation results as readily as a department's.
Corporate Identity on Login Pages
The logo and website icon defined under Settings > Other > Corporate Identity now apply to login pages as well, making branding consistent from the first screen a user sees. A definition made at host level reaches the login pages of every tenant, while a tenant's own definition overrides it for that tenant only. The host login page, shown before a tenant is known, uses the host-level definition.
Transcript Access Without Analytics
The Transcript tab on a conversation is no longer tied to the Analytics feature alone. It is now shown when any of Analytics, Agent Assist or Virtual Translator is enabled for the tenant, so customers who have licensed only Agent Assist or Virtual Translator can review conversations and their transcripts on the platform.
Improvements
Longer Agent Names
The Name and Surname fields now accept up to 200 characters, raised from 64, both on the user creation and edit form under Administration > Identity Management > Users and on the external API endpoints that create and update agents. Integrations that previously failed on long names now succeed, and a value beyond the limit returns a validation message stating the permitted range and the length supplied. The two-character minimum is unchanged.
Department Matching in the Virtual Agent Integration
When the department and agent synchronisation endpoint receives a department whose name matches an existing one but whose external identifier differs, the existing department's identifier is now updated instead of the request failing. Agents in the same request are then attached to the updated department. This keeps department and agent matching intact while the Virtual Agent integration moves from project identifiers to project names. Departments with no name match continue to be created as before.
Documentation Links Hidden in Whitelabel Mode
With whitelabel mode enabled, the links to product documentation in the settings screens are now hidden together with the sentences carrying them: the "You can find the full list here." links under the AI insights language, System language and Languages settings, and the "policy restrictions" link on the licensing settings. When whitelabel mode is off, all of them appear as before.
Predefined Categories Removed from New Installations
The "Casual Language" and "Negative Expressions" categories, together with their Turkish, English and Arabic variants, are no longer created when a new tenant is set up, and the evaluation form questions based on them are no longer included in the predefined forms. Existing tenants keep their categories, form questions and evaluation data unchanged. The matching entries were removed from the category template list in version 5.34.
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.