A personalized navigation system (PerNav) refers to a system that helps people with vision problems in their daily life activities with a chatbot system. Many researchers are continuously working in this field to minimize such people’s dependability with the help of devices and methods like Ultrasonic systems, cameras, IR, and many similar technologies. However, every system has some limitations like ultrasonic are suitable for navigation but fails in object detection and camera-based technologies fail in low light. Also, they all fail at a common aspect of trust-building. Blind peoples can easily be a victim of fraud. Our proposed model combines IoT and deep-learning-based models connected through a chatbot to prevent such people from fraud. The main idea is to use object detection and face recognition to make people aware of their interactions. In this system input from the camera are feeded to different models for object detection, face detection and optical character recognition and the output of all three are used to generate a sentence that will be passed from a bone conduction unit to convey the messages to the impaired peoples. With the help of transfer learning now the models can be customized over manually collected dataset.