Aims: The study aimed to investigate whether cortical auditory evoked potential (CAEP) measures could be used to optimise middle ear implant (MEI) and bone conduction implant (BCI) fitting, with the goal of improving hearing outcomes in adults. Methodology: CAEPs were recorded from electrodes Cz and Fz in response to three speech tokens (/U/, /A/ and /SCH/) presented in sound field. Decision (response present or response absent) was assisted by statistical criterions (Hotelling’s T2 statistical test) and confirmed by visual inspection of the waveforms. Optimisation was performed only when current clinical setting did not elicit a CAEP response. Hearing outcomes were measured as pre- and post-optimisation pure tone averaged aided thresholds (PTA4) as well as adaptive speech in noise scores. Results: Of the 14 participants, three did not need optimisation, two could not be optimised, and seven were successfully optimised using CAEP measurement. Univariate and multivariate analysis showed significant differences pre- and post-optimisation for middle and high frequency (i.e., /A/ and /SCH/) tokens. No difference was recorded for the low frequency speech token /U/. Optimised settings resulted in an overall improved speech perception in noise score (M(SD) = 5.11(4.24) SNR) and an improvement in PTA4 (M(SD) = 11.25(7.9) dB) pre- minus post-optimisation. An overall high degree of satisfaction with the procedure was observed. Conclusions: These preliminary findings demonstrated that middle and high frequency tokens could be successfully optimised using CAEPs, resulting on significant improvement in hearing performance. Our results provide preliminary evidence supporting the use of CAEPs for optimisation of MEI and BCI adult users’ fitting.