Abstract: To overcome the limitations of traditional meteorological nowcasting monitoring and forecasting , and enhance the refinement and intelligence levels of nowcasting operations , this paper integrates multiple sources of data and develops a nowcasting warning product of heavy rainfall . This product enables objective guidance for nowcasting warnings of heavy rainfall with a refinement level down to the township level and real-time rolling updates at minutely level . Additionally , a chatbot is developed to achieve real-time and intelligent distribution of the product. This system has been deployed in the actual operations of the Guizhou provincial meteorological department and has been promoted for use in meteorological departments at the provincial , municipal , and county levels . Compared to individual cases , the release time of this product is , on average , 17 min ahead of the actual release time of rainstorm warning signals . When compared to the test results of warning signals in the past three years , the lead time of rainstorm warning signals across the province has increased by 16 min and the hit rate has increased by 0 . 12% after the deployment of the system product in 2024 and under the combined effect of objective factors such as the response of duty forecasters . The nowcasting warning product for heavy rainfall and the chatbot developed in this paper have improved the refinement and intelligence levels of nowcasting operations , which is an important support in the warning release operations of grassroots meteorological stations .