cloudchat - Feature Request

Context: We are currently able to view the number of tickets assigned per agent through the platform's reports. However, there is no way to identify at what time these assignments took place throughout the day. This limitation makes it difficult to monitor operationally and manage team productivity in real time. We would like to view: Exact time at which each ticket was assigned to the agent; Number of tickets assigned per time slot; Evolution of attributions throughout the day; Comparison between agents considering the assignment time. Practical example: Today it is possible to know that one agent received 30 tickets during the day and another received 10. However, we were unable to identify whether: The agent started taking tickets right at the start of the shift or a long time later; There were long periods without new assignments; Productivity was constant throughout the day; There is an imbalance in the distribution of workload among agents. To answer these questions, we need to perform manual analysis and cross-reference information outside the platform. Suggestion for improvement: Add an assignment view with temporal detail, allowing you to monitor: Time of each assignment; Number of assignments per hour; Productivity timeline per agent; Filters by period and agent; Export of this data for analysis. Benefits: Greater visibility of the operation in real time; Better monitoring of individual productivity; Quick identification of delays in the start of services; Support for managing scales and team sizing; Lower dependence on manual analyses on spreadsheets.

None of the available APIs return the company of the contact linked to the call, making it difficult to analyze larger customers who offended the support

Today, the ordering of conversations is a standard user property: when defining an ordering criterion (e.g., most recent, SLA, etc.), it is applied equally to all filters. The request is to allow saving a specific ordering for each filter created, so that each filter maintains its own ordering criteria persistently, regardless of the others. Thus, when switching between filters, each one would respect the ordering configured for it.

When downloading an attached PDF from CloudChat, the file is saved with an extra suffix -filename_= at the end of the name. Example of the resulting name: document_example.pdf-filename_= . As a result, the operating system doesn't recognize the file as a PDF — the type appears as PDF-FILENAME_= File and doesn't open directly in the standard PDF reader. To use the file, it is necessary to rename it manually by removing the suffix. In operations with a lot of daily downloads, this manual renaming becomes an operational bottleneck. Expected behavior: file downloaded with the original name and the correct .pdf extension, opening directly in the standard PDF reader without the need for manual renaming.

Today, every time a ticket is solved, the CSAT is submitted. It should be possible to configure what types of tickets the CSAT is sent on. --- Consolidated from the post “Filter for submitting CSAT” : Allow the blocking of the sending of CSAT in certain calls (e.g., cases of spam). Today there is no option to filter which interactions receive the survey or not. Request: https://cloudhumans.unthread.io/dashboard/inbox/me/conversations/a3451ae7-65f7-4917-9fee-c04f9e53a13a

Today, when manually snoozing a conversation, the agent can choose the duration — “until the next reply”, “1 hour”, “tomorrow”, “next week” or a custom date. When the same snooze is configured as an action within a macro, the time option does not appear in the constructor: the snooze_conversation action is selectable, but there is no field to define how long the conversation will be delayed. Request: equate the behavior of the macro to manual snooze, exposing a duration selector (next answer/1h/tomorrow/next week/custom date) in the configuration of the snooze_conversation action within the macro. Motivation: macros are used to standardize repetitive service flows. Without the duration, macro snooze is tied to the standard behavior and forces the agent to snooze manually afterwards - which negates the macro's productivity gain.

Package of UX/UI improvements to the web widget, mapped based on internal use and customer request. Tooltip on the end conversation button: add tooltip to the end conversation button/icon, making the action clear when hovering. Today there is no textual indication of what the button does. Border of the text field following the color of the widget: the border surrounding the text field should use the same primary color configured for the web widget, maintaining consistency with the selected visual identity. Today the color of the border differs from the color of the widget. Visibility when typing messages with more than one line: when writing a message that exceeds one line, the user loses visibility of what they are typing. The field should automatically expand/scroll to keep the text always visible. Customizable “Connected AI” message: When the AI is connected, the response arrives within seconds, but the standard message doesn't reflect this and isn't editable. Convenia has already requested that this text be open/editable - it makes more sense to make it customizable per customer. Text translation according to the language selected in the input box: the text must be translated according to the language selected in the input box (today it is now possible to select the language in the input box). Automatic title/summary of the request: allow the title in a simple way - via AI that automatically summarizes the request, or the customer fills in the title when opening the conversation, or the agent completes it.

During the agent's response, the screen is static for 30—40 seconds without any visual sign that processing is in progress. This creates anxiety in the end customer, who does not know if the message arrived, if the system crashed or if it is worth waiting. Request: include visual feedback during this interval. Examples: “Typing...” indicator (messaging pattern). Temporary message such as “I'm looking at the information, one moment”. Motivation: reduce abandonment and complaints in lengthy processing windows, aligning customer perception with actual system behavior.

Example: an agent has access to edit the knowledge base but doesn't necessarily have administrator access

It will be possible to search Help Center articles through an internal search bar (within CloudChat), and not just through the external link. --- Consolidated from the post “A search tab when creating/editing the Help Center” : Even separating articles into categories, it is still difficult to locate specific articles - for example, in a category with 150 articles, finding one takes a long time. A search is also missing in the creation/editing area.