The efficiency of collision-avoidance abrupt braking or manoeuvring is primarily based on a driver's response time. The latter is affected by the driver's spatial and situational awareness, which in turn is heavily depended on the driver's cognitive workload. Attention taxing, infotainment systems could dramatically reduce the driver's ability to respond effectively in an imminent collision situation. Current attempts to reduce this negative impact on driver's performance had limited success. To improve the driver's ability to perform a successful collision avoidance braking or manoeuvring, this paper presents the design considerations of a prototype system that employs Augmented Reality (AR) to overlay guidance information in the real-life environment. The proposed system will be further supported by an Artificial Intelligence (AI) system that will act as a co-driver, offering in realtime alternative options to the driver. Prior work for the development of a similar system for Emergency Services' (ES) vehicles sparked the idea to transfer and investigate the acceptance of this technology on a civilian vehicle domain. In conclusion, the paper presents the preliminary design for the development of the civilian version of the AR/AI system based on the feedback and suggestions of twenty drivers.