Extreme wildfires are increasingly hazardous, particularly plume-dominated fires, which exhibit unpredictable behavior due to their self-sustaining convective columns. Despite their significance, these fires remain poorly understood, hindered by limited observational data and fragmented remote sensing approaches. This paper proposes an Artificial Intelligence-based method adaptive framework for real-time detection and characterization of plume-dominated wildfires using satellite and aerial imagery. The framework focuses on three critical aspects: fire intensity, vertical plume development, and rotational motion. Leveraging satellite data, alongside aerial datasets, the methodology dynamically adapts to available data to ensure accurate assessments. The final output is a risk-graded map identifying zones of active or potential plume-dominated activity, and a deeper characterization of the plume internal mechanisms. The proposed framework has significant potential for improving wildfire prediction, management, and mitigation strategies, contributing to improved safety and resource allocation in wildfire-prone regions.