Occasionally, contacts might be replying to a specific message that the contact has sent or the AI Agent has sent. In this case, the replied-to message does not get passed to AI Agent as context. This may lead to hallucinations as AI Agent does not have context of what the contact is replying to.

Similar to ChatGPT, we can furnish the messages that AI Agents received with the message that the contact is replying to. This should provide AI Agent with more context on what the conversation is about, and hence AI Agent's reply can be more focused and accurate regarding the context.