A method, system and computer program product for providing an automatic cueing system for real-time communication, by: collecting a first person's profile data, which may include a medical history, to determine how the first person reacts to different voice or audio characteristics and sounds； analyzing the first person's voice or audio characteristics and sounds to categorize a current condition of the first person； and then dynamically adjusting voice or audio characteristics and sounds of responses made to the first person by a robotic device or a second person, based on the first person's profile data and the current condition of the first person. Thereafter, the first person's voice or audio characteristics and sounds, as well as the responses made to the first person, may be continuously monitored, along the first person's vital signs, to update the current condition of the first person.
