In this study vector quantization (VQ) was used as an efficient vehicle to implement a simple adaptation process integrated into an isolated word, speaker-independent ASR system, based on dynamic time warping and linear predictive coding. VQ indices computed from incoming utterances are translated on a one-to-one basis by the personalised map prior to entering the recognition stage. Assessment of adaptation is given using Grenier's ratio, for which the system has a maximum ratio of 0.82.