This paper explores the application of generative artificial intelligence as a coding assistant to produce land use and land cover transition maps within the Google Earth Engine (GEE) platform. Focusing on urban expansion in the State of São Paulo from 1985 to 2024, we propose a protocol that enables researchers—particularly those with limited programming expertise—to harness the power of AI to utilize GEE's extensive geoprocessing capabilities. The protocol concurrently implements quality assurance layers to guarantee the accuracy of the final map product. The successful results of the mapping demonstrate the viability and effectiveness of this approach. Portuguese version: https://doi.org/10.1590/SciELOPreprints.13795