Separate "Stop/Start AI Bot" and "AI Feedback Controls" from "Manage AI Settings" Permissions

Description

The Problem:

Currently, there is no way to allow an agent to manage a live conversation without exposing the bot’s configuration. To let an agent see the "Stop/Start Bot" button or the AI Feedback icons (thumbs up/down, star for source info), we must grant them broad "AI Agent" permissions.

This creates two major issues:

Security/Privacy Risk: Through the "Star" icon (source info), any employee can trace back the response and, in many cases, navigate to the AI Prompt or settings. Even if the "AI Agents" tab is hidden via custom CSS, these UI elements provide a "backdoor" to the bot's logic.

Data Integrity (Knowledge Base Pollution): Employees often click the "Thumbs Up" or "Thumbs Down" without proper training or by mistake. This automatically trains the bot with potentially incorrect data, "polluting" the Knowledge Base and making it difficult for agency owners to maintain a clean and accurate AI response system.

Proposed Solution:

We need granular, standalone toggle permissions in the User Permissions settings:

[Toggle] AI Operational Controls: Allows the user to see and use the "Stop/Resume Bot" button in the conversation window ONLY.

[Toggle] AI Feedback & Insights: Allows the user to see the Thumbs Up/Down/Star icons. If disabled, these elements should be completely hidden from the UI for that user.

Benefits:

Scalability for Agencies: We can onboard client staff to handle chats without fearing they will accidentally modify the prompt or corrupt the AI training data.

Clean Data: Only authorized/trained managers would be able to "upvote" responses, ensuring the Knowledge Base remains high-quality.

Better UX: Non-admin users get a cleaner interface without "noisy" buttons they don't need to use.

Current Workarounds (and why they fail):

Using tags to "pause" bots is inconsistent due to browser glitches or latency. Using CSS to hide menu items is not a security solution, as the AI feedback icons still allow access to the bot's origin and settings.