<p>This paper summarizes the research work that was completed to (1) gain an in-depth understanding of pavement temperature differentials during construction through thermal profile data, (2) apply innovative machine learning techniques to find the optimal use of available thermal profile data, and (3) develop a fair and executable specification for employing Paver Mounted Thermal Profiling (PMTP) technology to ensure high-quality asphalt pavement construction. While PMTP technology shows promise for continuously detecting pavement thermal segregation, widespread adoption of pavement quality assurance protocols is still nascent. PMTP data from 24 pilot projects spanning 2018–2023 were analyzed using advanced machine learning clustering algorithms. This method was used to grade the commonly used measures for quantifying pavement thermal segregation and then develop a specification recommendation based on the best metric. Using ArcGIS Pro software, this research was able to quantify pavement thermal segregation on a granular level using the K-Means++ clustering algorithm and geospatial post-processing. Results show that the differential range statistic (DRS) is 26% better at measuring thermal segregation when compared with the thermal segregation index (TSI). A specification recommendation was developed by optimizing the relationships between DRS payments and existing density-based payment factors. These payment factors, weighted by the quantity of paving materials, tied the specification to performance. The recommendation for implementing DRS is to use a payment penalty to incentive ratio of − 0.40 that aligns with existing specifications and provides options for further specification development using the advanced clustering methods applied in this research.</p>