P
Pasqual Toral
It would be very necessary to separate the configuration related to the prompt, behavior and all these types of technical functions, from the chat switch that allows the agent to activate, deactivate or put to sleep for a while.
I will explain two situations where this is interesting:
- Global context:
Let's consider an agency that offers GHL to other companies, where we are going to consider 3 people:
_Agency User: Employee or administrator of the company that acts as a service-providing agency and who has full control over customer accounts.
_Administrator Customer: Person responsible for managing GHL within your company, whether someone from IT or the boss. It would have the GHL Account - Admin user profile type, thus the highest level within the customer's sub-account.
_Customer Employee: Any other employee in the company who only uses GHL as a day-to-day tool should generally not have access to any configuration.
- Situation 1: (Our particular case)
The agency offers GHL as a closed solution to its clients and managed primarily only by an Agency User. Any major changes are consulted with the Agency and the changes are made with a draft amendment, to avoid failures.
Customer Administrator only has access to basic administration: view employees, some specific configuration function...
The Customer's employee only uses the tool to keep the agenda, respond to customers and little else.
_What are you looking for?
In this case, the Agency, everything that is AI, Automations, technical configurations... does not want either the Managing Client or the Client's Employee to be able to see it, so that they don't break anything. And it wants to implement an AI Conversation Agent in the Customer's sub-account to reply to messages. He wants at least the Administrator Client to have access to the button to turn on, turn off or put the AI Agent to sleep from the Conversations panel, but to nothing else, so that the promt configuration cannot be loaded. And from there, the Managing Customer already chooses whether or not the Customer's Employee can see that button.
- Situation 2: (A more generic case)
Customer Administrator has access to everything and creates AI chats for use by their Customer Employees. The Agency here gives access to full functionality to the Managing Client and he himself manages the permissions for his Client Employees.
_What are you looking for?
He has created an AI for customer service and it is activated for employees, but he only wants them to be able to stop the AI in chats when appropriate but not have access to anything else, since for example they are only the secretary or receptionist or any other employee who should not have access to anything technical so as not to break it.
(This situation also applies at the agency level, an agency may have employees and want them to serve people, and thus be able to, for example, deactivate AI in a client who does not want AI to serve them, or pause AI while they are having a conversation. But you clearly don't want those people to be able to touch anything in the settings since they're just customer service employees, with no technical profile, and they could break something.)
- Global Problem
The configuration does not distinguish the button that allows you to turn on, turn off or put to sleep (I think it's called Change of State button) from the rest of the configuration, (henceforth, “the Button”) it only allows you to enable or disable “View and manage conversational AI agents”, which encompasses the entire configuration and chat button.
  • In situation 1: This means that in order for the Managing Client to use that button, the internal configuration of the chat must also be shown to him, and he can even delete the agent or deconfigure any function. And the same thing happens to him if he wants to allow Customer Employees to turn on and off the AI for each chat, they will have to see all the internal configuration and be able to break everything.
  • In situation 2: If the Managing Client wants his employees to be able to deactivate AI for a specific chat, he must allow them to see the internal configuration of the chat and thus be able to load the AI Agent. And if he doesn't want to allow that, then he has to hide everything related to the agents and tell his employees to let him know if anything needs to be changed, having to personally manage any activation or deactivation of the AI.
- MUCH NEEDED SOLUTION
A clear separation of these two parts of the functionality of the Conversations AI Agent, which are the Button in the conversations interface and the internal configuration of the agent, and which can be configured per user in each sub-account. This will ensure the security of the configuration of agents in clients.
This is a very basic functionality of agent systems and should be native to the Conversation AI Agent configurations.
  • References:
Old related request: https://ideas.gohighlevel.com/conversation-ai/p/granular-permission-for-stop-start-ai-bot-in-conversations
There was never an answer and it mixed other features that are now there, so I made it new.
·