This paper considers the ascent-to-slot problem of an electrically propelled geostationary satellite. In contrast to the widely-studied electric orbit transfer problem, the present work is concerned about the rendezvous ascent problem, targeting a specific geocentric longitude. This problem has been less frequently dealt with in the open literature. An artificial neuralnetwork-assisted fast and efficient algorithm for generating such trajectories is introduced, which can easily subsume arbitrarily high-fidelity force models and accommodate eclipses. The simulated ascent trajectories, starting from a variety of injection orbits, demonstrate the efficacy of the presented method. The proposed algorithm, and its software implementation, can be a valuable addition to flight dynamics analysis tools for all-electric satellites because of its potential to enable automated exploration of low-thrust mission trades.