In past many researchers have designed the multiple robots which served as per the need of humans but unfortunately they are much expensive. It has been observed that man power is needed in the shopping marts especially on highway based stores and marts. Where it is found to be inconvenient for persons or owners to perform the duties all the time. This paper is centered around one of the best solution in this regards that is the cheapest man–machine association by discourse where the robot can perceive the discourse and it can react to the individual in that instant. In this paper the prototype of a speech recognition based robotic mart has been designed and analyzed which is based on speech recognition system. To make it semi-autonomous it compares the data in database unit, if the data matches with any command within the database, it instructs to main system to perform that function accordingly. The proposed robotic mart is supposed to be an economical and advanced robotic system that consists of cheapest Arduino microcontroller as a main control unit and a smartphone based application software. Its GUI framework will help the user for addressing the gadget; it will carry the accommodation to the user discussing voice worked gadget controlling hence, decreases the human endeavors. User doesn't need to get up from his place to control the device and user can talk in mobile or some other gadgets utilizing the MIT application stage and afterward the directions are sent to the control unit through remote correspondence procedure by means of Bluetooth.