Pneumonia is one of the major causes of death, especially in paediatric groups, with late diagnosis being a major risk factor that complicates the clinical situation and treatment. Traditional diagnostic methods based on the use of manual decoding of chest X-rays are limited by inter-observer error and the lack of specialists, especially in high-volume or resource-intensive environments. This paper suggests a vision-based detection system based on a convolutional neural network that was trained on 5,863 chest X-ray images to classify pneumonia (binary). The framework is also not limited to diagnosis, it incorporates a simulated robotic intervention module that can invoke automated alerts, clinical notification, and preliminary response activities. The model uses standardised preprocessing and has an efficient feature extraction with about 11.17 million parameters. The system has a test accuracy of 89.74% with a recall of about 96.7%, which means that the system is highly sensitive in detecting pneumonia cases. Combining AI-based detection with robotic action shows that a scalable solution can be used in real-time clinical assistance, especially in intelligent and remote healthcare settings.