In this paper, we propose a novel integrated sensing and communication (ISAC) complex convolution neural network (CNN) CSI enhancer for 6G networks, which exploits the correlation between the sensing parameters, such as angle-of-arrival (AoA) and range, and the channel state information (CSI) to significantly improve the CSI estimation accuracy and further enhance the sensing accuracy. The ISAC complex CNN CSI enhancer uses the complex-value computation layers to form the CNN to better maintain the phase information of CSI. Furthermore, we incorporate the ISAC transform modules into the CNN enhancer to transform the CSI into the sparse angle-delay domain, which can be treated as images with prominent peaks and are suitable to be processed by CNN. Then, we further propose a novel biased FFT-based sensing scheme, where we actively add known phase bias terms to the original CSI to generate multiple estimation results using a simple FFT-based sensing method, and we finally calculate the average of all the debiased sensing results to obtain more accurate range estimates. The extensive simulation results show that the ISAC complex CNN CSI enhancer can converge within 30 training epochs. Its CSI estimation normalized mean square error (NMSE) is about 17 dB lower than the MMSE method, and the bit error rate (BER) of demodulation using the enhanced CSI approaches the perfect CSI. Finally, the range estimation MSE of the proposed biased FFT-based sensing method can approach the subspace-based method with much lower complexity.