Public service provision in the frontline, coined street-level bureaucracy, has been gradually impacted by information and communications technology (ICT) for decades. This impact, however, has mostly considered ICT as a tool suitable for tasks with low complexity. With recent advances in artificial intelligence (AI), there are examples of AI use for more complex street-level work. Examples include cases where AI is used for assessing eligibility for social benefits, predictive policing models, automated grading, and diagnostics in healthcare. While these applications demonstrate potential benefits, they also introduce new challenges related to privacy, accountability, corporatization and alienation of street-level work, and client service experiences. This article is a critical reflection on the street-level potential of AI in providing public services. This study contributes to the ongoing debate on AI's impact in street-level work by emphasizing both the potential benefits and risks associated with AI integration in frontline service provision. While AI may mitigate some limitations of human decision-making (e.g., subjectivity, inconsistency, and bias), it can also introduce challenges that require careful consideration (e.g., lack of transparency, data-driven bias, and limited contextual adaptation). By critically reflecting on AI's street-level potential, this article calls for a balanced approach to AI adoption in street-level work.