This study demonstrates how raw sewage can be used as a practical surveillance tool to track antimicrobial resistance (AMR) trends at wastewater treatment works. Over 24 months, samples from four WWTWs in Gauteng were analysed using metagenomics, enabling AMR signals to be measured without relying only on clinical isolates. Results showed that aminoglycoside, beta-lactam, sulfonamide and tetracycline resistance were consistently most abundant across sites, with differences observed between municipalities and between plants serving different socioeconomic catchments. Seasonal analysis showed a general trend towards higher AMR gene abundance in colder seasons at certain sites. These findings support the value of wastewater surveillance as an early-warning layer that can complement existing public health and environmental monitoring programmes.