• workshop_dense_100#slide 32

Within the RAIDT programme, the Theory & Foundation branch provides the conceptual spine that makes later empirical validation and policy translation cumulative rather than fragmented. The academic-logic paper shows the architecture clearly: the foundations paper defines RAIDT itself, while feeder papers supply bounded theoretical work on design science, socio-technical significance, governance theory, audit lineage, and technical grounding. Among the papers cited here, 12 carries the design-science and mechanism-based mid-range design theory argument, 17 explains the organisational and managerial significance of reviewable run-level governance, and 11 integrates these strands into a PhD-level statement of problem, gap, method, evidence, and contribution.

Substantively, the branch starts from a precise residual gap. Existing principle sets, documentation artefacts, explainability techniques, and audit routines do not converge on a standard proof object for one configured generative AI use in organisational work. The branch therefore theorises run as the unit of governance and defines the core RAIDT artefacts: the run-level evidence pack and the score profile across the five pillars (Responsibility, Auditability, Interpretability, Dependability, Traceability). It also clarifies that governance quality must be specified through inspectable evidence, not narrative assurance, and rendered through anchors 1=missing / 3=partial / 5=audit-ready.

The branch matters because it does three jobs at once. It explains why RAIDT is a legitimate design science research contribution, why run-level evidence changes socio-technical governance practice, and how influence methods as governance interventions can be studied systematically. In short, this branch supplies the theory, artefact logic, and organisational meaning that the rest of the RAIDT programme depends upon.

In a public-service eligibility workflow, a local authority uses generative AI to draft benefit guidance. The Theory & Foundation branch would not start by asking only whether the model is accurate in general. It would ask what evidence object allows the council, an auditor, and a challenged citizen to inspect this specific use. The answer is a run-level evidence pack linked to a score profile.

Here the branch's internal division of labour becomes visible. The design-theory component explains why this is an artefact problem; the socio-technical component explains how managers, compliance teams, and reviewers use the artefact; and the academic-logic component shows how the resulting knowledge contribution hangs together across papers. The council can then compare runs, sample weak cases, and improve governance routines without collapsing back into policy slogans or ad hoc logging.

• 11-RAIDT_Academic_Logic_M_v11
• 12-RAIDT_DSR_Theory_M_v8
• 17-RAIDT-Sociotechnical_M_v6