The current problem seems to be not the lack of high-quality coastal morphodynamics measurements but the ability to analyze vast amounts of such data. Current GIS methods are insufficient, so developing newer, more automated ones is necessary. The study aims to test the current automation potential in coastal morphodynamics monitoring. The tideless shore of the sandy barrier of the Hel Peninsula (Gdańsk Bay, southern Baltic Sea) was examined using digital elevation models (DEMs) obtained from monitoring the Polish shoreline. DEMs were analyzed using the Beachmeter tool in the ESRI ArcGIS Pro. The Beachmeter delineated the beach area, calculated its basic morphological parameters for each measurement campaign, and determined the sediment budget. The analyses were performed semi-automatically significantly faster than manually. The difference maps generated by Beachmeter were then automatically interpreted using the developed GeoAI model. A U-Net convolutional neural network was designed and trained to segment these difference maps. This deep learning-based approach automatically identified areas of significant erosion and accumulation. Deployed experimental coastal monitoring methods were assessed as valuable and worth further development. The advancement of geospatial artificial intelligence should be harnessed in coastal geomorphology, supporting researchers in their investigations and opening up new opportunities for spatial data analysis. Simultaneously, further development of automated tools for coastal monitoring is necessary, although semi-automated beach morphodynamics monitoring is possible today, as this study demonstrated.