This study examines the dynamics connected to the provision of internet service in Tanzania. Specifically, the study seeks to systematically analyse user perceptions service quality across different ISPs and identify the factors influencing their preferences and choice of ISPs. Two theoretical models, the Service Quality Theory (SERVQUAL) and the Expectancy Disconfirmation Theory (EDT) were applied to guide the study. A qualitative single case study design was used, with more than 800 user-generated comments extracted from a celebrity’s Instagram post. Data were analysed using sentiment analysis, qualitative and quantitative techniques. Our results show that while users acknowledge improvements in reliability of network, still inconsistencies remain, particularly in specific zones of Dar es Salaam and rural areas. T-CONNECT is perceived as having the most reliable network compared to other ISPs. Speed consistency remains a critical issue for all ISPs, with T-CONNECT seen as the most stable, whereas SKT-TZ and TZ-NET struggle, especially during peak hours. SKY-TZ, TZ-NET and TZ-LINK face coverage challenges, especially in rural areas due to infrastructure constraints. Moreover, the study highlights a high demand for reliable service, a push for fiber-optic infrastructure, and a cautious but optimistic outlook on 5G. Generally, the findings reveal that customers’ perceptions of internet service quality are influenced by network quality, and customer service and technical support. Challenges such as connectivity and network issues, pricing transparency and service plans, and customer support significantly affect user satisfaction and ISPs competitiveness. The contribution of this paper stems from its use of advanced NLP methods to examine user sentiments toward Tanzania’s ISPs from Instagram data, to provide insights and comprehensive understanding to industry practitioners and policymakers in promoting a more sustainable future for the industry.