From decision support to decision infrastructure: a digital decision systems research agenda for the operational research and information systems interface

Digital systems no longer merely inform operational decisions; they construct rosters, set prices, match participants, triage exceptions, and execute allocations at scale. Yet formally sophisticated models do not reliably produce superior operational outcomes, and widely used systems can generate brittle, inequitable, or resisted decisions: neither model quality nor technology use explains realised decision value. This paper argues that the research object capable of providing that explanation is the digital decision system: a sociotechnical configuration in which data, models, algorithms, human discretion, governance, and organisational context jointly produce operationally consequential choices. We define the construct, distinguish it from adjacent notions, and develop a relational architecture and a two-level temporal model linking a recurrent decision cycle to a longer system lifecycle. On this foundation, we build a mechanism-based research agenda organised around four causal problem domains (problem representation, prescription–workflow coupling, discretion and exception governance, and adaptation), each with orienting propositions, research questions, and methodological pairings, together with boundary conditions and contingent predictions. The agenda gives the operational research and information systems interface a shared research object and a cumulative program, and it repositions operational research as the discipline best placed to theorise the decision logic of digitally mediated operations.