Security is one of the possible fields in human society in which robotics can be applied. Human guards usually perform a range of tasks in which a robot can provide help. A security company collaborated with us in the design and development of a robot that should serve in patrolling large indoor areas, interacting with humans, welcoming, providing information, and be a telepresence platform for the human security guards. In this paper we present a preliminary experiment which involved this new robot in two roles: security and guidance. The former being important especially during night, and the latter being common in daytime, when guards are usually interacting with people who ask them information. The results of the experiment with 55 participants showed how the perception of the appearance of the robot and its effectiveness are influenced by its behaviour and its related perceived more authoritative or kinder traits. These results provide useful indication for the employment of robot guards in a real world situation.