Gravel road condition assessment is crucial for effective infrastructure maintenance. Visual windshield surveys, the traditional assessment method, are often inefficient and subjective, leading to inconsistencies in maintenance decisions. This paper presents the preliminary results of a study that tests different methods for assessing the condition of gravel roads and explores possibility of integrating these assessment methods for improved effectiveness. Field experiments were conducted on a gravel road directly after maintenance and several months later after corrugations had formed. Visual assessments and the International Roughness Index (IRI) measurements were performed using the Roadroid smartphone application. At the same time, the vehicle vibration response data was collected using an Integrated Electronics Piezo-Electric (IEPE) accelerometer. The findings demonstrate a correlation between higher IRI values, increased vehicle vibration levels and the presence of corrugations. Analysing vehicle vibration responses in different frequency sub-bands also provided insights into specific road defects, such as loose gravel and corrugations. Integrating these data sources provides a more accurate and detailed insight into the gravel road conditions that would enable targeted maintenance interventions. There is potential to incorporate artificial intelligence (AI) for automated road condition assessment and predictive maintenance, for instance, through a machine learning model. The model trained on the collected data would then assess the gravel road condition and predict future deterioration, thus optimising maintenance resource allocation and improving efficiency and cost-effectiveness in managing gravel roads.