I appreciate the powerful automation tools ManyChat provides, especially the ability to access the last_user_input variable. This functionality enables seamless integration with external tools, such as an AI assistant, by transmitting the user's last message.
However, in real-world scenarios, live chat agents frequently step into conversations between users and the chatbot. Currently, there is no built-in way to distinguish between messages sent by a bot and those sent by a live chat agent. This presents a challenge when integrating with AI assistants, as they may lose critical conversation context when automation resumes.
Feature Request: I propose adding two new variables:
-
last_agent_input– Capturing the last message sent by a live chat agent. -
last_agent_input_timestamp– Storing the date and time of the last message from a live chat agent. (requested also here)
Benefits:
-
Enhanced AI Training & Adaptation: AI assistants, such as GPT-based integrations, could learn from agent responses, which are often superior to bot-generated messages. This would help improve chatbot responses over time.
-
Maintaining Conversation Context: When automation is paused while an agent handles a conversation, the AI assistant should still receive and process the agent’s messages. This ensures the AI retains full context once automation resumes.
-
Seamless AI-Assisted Live Support: By forwarding agent messages to an AI assistant, businesses can refine responses, analyze customer interactions, and improve chatbot accuracy based on real human interactions.
By implementing these variables, ManyChat would empower users to build smarter, more context-aware AI integrations, ultimately improving the quality of automated conversations.
Thank you for considering this feature request. I believe it would be a valuable addition to ManyChat’s capabilities, helping businesses provide a more seamless and intelligent customer experience.