Feature Request

Please implement a connector with Odoo

Currently, if an agent ends their shift but remains “Online” for an open session, the system still considers it available and continues to incorporate it into the Round Robin. This generates an artificial assignment that directly impacts SLA, operational load and customer experience. Therefore, it makes perfect sense to incorporate specific functionality for administrative disconnection by roles with permissions (Supervisor/Administrator). Functional Proposal: “Forced Disconnection by Supervisor/Administrator” The idea is to enable a supervisory control that allows the agent to change the operational state remotely when it is detected that it was connected out of shift, immediately excluding it from the assignment engine. “Disconnect/Go Offline” action by authorized role. With one click, the supervisor/administrator executes an action that: 1- Change the agent's status to Offline/Not Available. 2- Remove it from the Round Robin immediately. 3- Avoid new assignments from that moment on.

Restricting agents from seeing unassigned conversations is fine, but they still have access to the option to start conversations from any channel in the bottom menu, which doesn't make sense.

Business Problem: As AI Agent becomes more capable, the reasoning behind their responses and actions become increasingly important. Reasoning is what AI uses to interpret user intent, decide next steps, and orchestrate actions across automations. Currently, reasoning is hidden. When an AI Agent sends a message or performs an action, users have no visibility into why it happened. This makes it difficult to: Troubleshoot unexpected AI behavior Validate that AI is acting in line with user intent Build trust in AI-driven conversations. Desired Outcome: Introduce a feature to expose AI Agent’s reasoning behind messages and actions sent to a Contact. This allows users to understand why AI responded or acted in a certain way, improving transparency, trust, and debuggability for teams using AI Agents.

Use case: that external integrations (CRM, backend, etc.) can update the lifecycle stage via MCP.

Add the ability to import a ZIP file with the same folder structure as the files that have been imported, Folders, Sub-folders.

When a lead lifecycle changes from new lead to junk/cold lead, the system requests a reason why this change to that state, along with the ability to select prefilled categories with a simple description, and this event is available to automate it and route it to your desired CRM.

For industries with a long sales cycle, it is not really applicable to close conversations.

Please add the ability to organize the attachment into very well-structured files and to be able to add subfolders too, along with choosing languages for files too, and showing the file structure in the inbox when you choose attach from libraries.

Currently, when a single contact starts conversations on multiple channels (e.g., WhatsApp + Facebook) around the same time, the workflow triggers only for the first channel. As a result, channel-specific AI agents for the other channels never activate. Desired behavior: Each time a contact starts a conversation on a channel, the workflow should trigger independently for that channel. This would allow the correct, channel-specific AI to take over smoothly, keeping conversations clean, contextual, and properly segmented. Why this matters: We actively use the unified inbox and plan to scale by adding more channels and multiple workspaces for sister concerns. This is not an edge case—it becomes a real limitation as volume and channel diversity grow. Without per-channel workflow triggering, AI routing breaks in multi-channel scenarios, reducing automation reliability and increasing manual intervention. This feature would significantly improve scalability, AI accuracy, and multi-brand/multi-channel operations for teams like ours at Betafore. Opinionated take: unified inbox without channel-aware triggers is like a traffic signal that only turns green once—great idea, awkward execution.