We present a study that uses an artificial intelligence algorithm (e.g., Fuzzy c-mean clustering) for the development of an alert system for commuters using air pollution monitoring data over a coastal city of eastern India. Analysis of inter-annual and inter-seasonal variability of air pollution was carried out over Bhubaneswar (85.81º E, 20.35º N, 48 amsl) during 2018–2019. The particulate matter of size 10 µm (PM10) observations from Envirotech APM Dust Sampler and the 7-day trajectory history from Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) archives were implemented in hybrid source-receptor models to estimate the potential of virtual source fields towards the pollution at the receptor. A novel approach of source-receptor analysis using the FCM was implemented to find out the aerosol emission source regions in terms of clusters of air-mass pathways, virtual source- and concentration fields. Results show that the ambient air pollution was observed to be as highest in winter, with values exceeding the permissible limit of annual average figures (100 µg m−3), whereas for the remaining three seasons the seasonal average falls below the permissible limit. Results show that transport of meso- to synoptic-scale level emissions towards the receptor station leads to severe pollution accumulation compared to that of long-range transport. In detail, the transport of air masses was observed from the Indo-Gangetic belt of India during winter (the prominent season as far as air pollution severity is concerned). The surface-level air pollution data was implemented in our study emphasizes to provide air pollution alerts to safeguard the commuters from exposure.