Summary Nowadays, most cochlear implant (CI) candidates have residual hearing, mostly in the low-frequency region. In these patients, peri- and postoperative preservation of the acoustic component is highly desirable. However, our current understanding of the residual cochlear function in CI candidates and recipients is insufficient. Moreover, we still do not fully comprehend how the inner ear reacts to the implant electrode. Therefore, we need to develop methods to improve our understanding in these matters. With the help of the implant electrode, we can record electrophysiological measurements (i.e., electrocochleography; ECochG) directly at the site of interest (i.e., in the cochlea). We can repeat measurements during and after implantation as often as required and thus reveal changes of the inner ear function. However, current analysis methods rely heavily on expert judgment. This makes the systematic investigation of intracochlear processes difficult or even impossible. We have introduced an objective assessment of ECochG recordings using a Deep Learning approach. The advantages of such objectification are as follows: the evaluation is standardized (allowing longitudinal comparison among individuals as well as comparison between different study centres), fast, accurate and examiner independent. By systematically collecting objective measurements of the inner ear, we can now reveal patterns of the inner ear function. This should increase our understanding of the residual inner ear function during and after implantation.