This paper presents a new in-car advisory system that gives advices on lane, speed, and headway. The advices are determined at a traffic management center based on a new lane level traffic state prediction model, in order to prevent or solve suboptimal traffic flow conditions. The system aims for an optimal lane distribution in high flow conditions, decreasing the chance of spillback by advising drivers away from the right lane, and a reduction in the capacity drop by advising drivers to maintain a short (but safe) headway at the end of congestion. The system is implemented in microscopic simulation to evaluate the potential benefits for different penetration and compliance rates. Benefits at both low and high rates are found as only a small redistribution of traffic over the lanes may be required to stabilize flow. The capacity drop is mainly reduced at high rates as it is required that many vehicles accelerate more. The maximum benefit found is a reduction of 49% in travel time delay. Effects are smaller at lower rates. Negative side effects are also found, including oversaturation of lanes partially by advised lane changes and increased probability of spillback taking effect.