Bringing Back the Front Desk: A Bounded Case for AI-First Public Services

Digital government has expanded access to public services, but it has also redistributed substantial interpretive and procedural work to residents. Many citizens and residents must now navigate fragmented portals, decode administrative language, identify the appropriate service pathway, and complete complex forms with limited direct support. These demands often fall most heavily on people with lower administrative literacy, limited language proficiency, fewer resources, or less confidence in dealing with public institutions. This article advances a bounded conceptual proposition: governments should examine whether artificial intelligence can function as the primary digital front door for selected categories of public services. The argument is not that conversational AI can remove the underlying complexity of law, eligibility rules, or institutional fragmentation. Rather, it may help residents navigate that complexity more effectively in some domains while also introducing new risks and burdens. I argue that AI-first public services are most plausible in high-volume, rules-rich, documentation-heavy, and relatively low-discretion interactions, rather than as a universal model for all citizen-state encounters. The article positions conversational government as a research, design, and governance agenda requiring staged pilots, domain-specific evaluation, evidence on burden redistribution, privacy safeguards, meaningful human oversight, and retained traditional support channels for residents who most need them.