The growing adoption of electric vehicles (EV s) as company cars significantly amplifies charging demand and produces large power flows that must be carefully controlled. The advent of bi-directional charging opens up opportunities while simultaneously adding layers of complexity to the design of charging plans. Concurrently, installing photovoltaic (PV) systems at the company premises enhances sustainability but also introduces fluctuations in power generation, posing challenges to effective energy management. These complexities are compounded by the inherent uncertainties in forecasting EV behavior, PV output, and energy pricing, which further complicate the design of smart charging solutions. This paper introduces a heuristic bi-directional smart charging algorithm that generates charging station assignments for limited bi-directional charging stations and creates charging plans in real-time while being robust to prediction error. The simulation results demonstrate that our approach yields considerable advantages compared to benchmarks, including cost reductions, peak demand management, and improved PV energy utilization.