You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Scenario:
Our project involves the development of an Autonomous Agent for Customer Care Centers. Acting as the initial point of contact for customers, this agent will autonomously handle generic inquiries and, based on user intent, delegate actions to the appropriate human personnel. The agent will also provide comprehensive summaries and manage all customer interactions.
Note:
The architecture we're implementing is event-driven and decoupled, facilitating seamless automation evolution without necessitating changes to the implementation. Transitioning from human to agent involvement is a straightforward process achieved by processing and dispatching the same events.
Key Points:
Dispatching will be executed using Cloud Events.
Events will contain parameters necessary for the agent to acquire through user querying, utilizing Language Model (LLM) capabilities.
Certain events will be flagged as "answer expected," prompting the agent to await a response event within the user session. Conversely, other events will be "fire and forget," with users notified through alternate means such as email.
We envision a "master Agent" endowed with the ability to dispatch tasks to both sub-agents and human agents, tailored to specific functions, including:
Updating customer information (e.g., address)
Investigating invoices
Providing discounts
Examples of sub-agents include:
User authentication (Agent)
QnA (AiAgent)
To-Do:
Define the scope of capabilities for both sub-agents and human agents.
No description provided.
The text was updated successfully, but these errors were encountered: