OBJECTIVES: To assess and compare medication counselling and dispensing practices at community pharmacies in three cities of Pakistan; Islamabad (national capital), Peshawar (the capital of Khyberpakhtoonkhwa province) and Lahore (the capital of Punjab province). METHODS: A total of 1113 patient-dispenser interactions were observed from a randomly selected sample of 371 pharmacies by using convenient sampling technique in the three respective cities namely Islamabad (n=118), Peshawar (n=120) and Lahore (n=133). The data collection tool was adapted from WHO structure observation form and was modified according to the objectives of the study. The data was coded, entered and analyzed by using SPSS Version 16.  RESULTS: Out of 1113 patient-dispenser interactions, the providers present at the community pharmacies were: pharmacist (n=18, 1.6%), pharmacy assistant (n=78, 7.0%), diploma holder (n=62, 5.6%) and salesmen (n=955, 85.8%).  Prescription validation was carried out in 18.5% of cases, drugs verification in 32.3% of cases and labelling in 6.8% of cases. In 40.3% of cases, dose was communicated, in 40.6% of cases frequency of medicine was told and in only 3.7% of cases, side effects were communicated to the patients/customers. Overall, only 35 (3.1%) patients were counselled completely about their medication, while 582 (52.7%) patients were not provided with any medication counselling. CONCLUSIONS: The process of medication counselling and dispensing practices at community pharmacies in Pakistan is not satisfactory. The patients are largely handled by unqualified salesmen who are also involved in product substitution which is beyond their work scope. Thus, there is a strong need to improve medication counselling and dispensing practices at community pharmacies which in turn will result in the provision of better patient oriented services at community pharmacies.