• integrated_82#Q2.12

RAIDT is framed as design science and mid-range design theory because the central research gap is artefactual. The papers show that principle-led governance, model cards, explainability methods, and audit routines are all useful, yet none by itself gives organisations a bounded proof object for one configured use of GenAI in context. That kind of gap calls for design science: the task is to specify, justify, and evaluate an artefact that can be used in practice. RAIDT answers that by treating the run as the unit of governance and by designing two linked artefacts, the run-level evidence pack and the score profile across the five pillars (Responsibility, Auditability, Interpretability, Dependability, Traceability).

The same contribution is formalised as mid-range design theory rather than grand theory because its claims are mechanism-based and conditional. RAIDT explains how evidence capture, scoring, and influence methods as governance interventions can improve or weaken governance readiness under stated organisational conditions. It is not a universal theory of institutions, regulation, or all AI systems. Nor is it merely a toolkit. The anchors 1=missing / 3=partial / 5=audit-ready make the artefact operational, while the design-theory framing makes it falsifiable and cumulative: researchers can test whether different configurations, review routines, or sectors produce different run-level outcomes, instead of relying on narrative assurance.

In a public-service eligibility workflow, staff may use GenAI to draft an explanation to a claimant about why support was approved, refused, or escalated. A grand theory of governance would be too general to tell the organisation what evidence must exist for that one message. Design science is more suitable because the agency needs a working artefact: a run-level evidence pack containing the exact prompt version, policy clause used, retrieval snapshot, output, and reviewer decision.

Once that artefact exists, RAIDT?s mid-range design theory becomes useful. The agency can see whether retrieval augmentation improved Traceability, whether reviewer sign-off strengthened Responsibility, and whether the overall score profile remained only partial because Auditability evidence was incomplete. That is actionable knowledge about one class of governance problem, not an abstract theory of all public administration.

• 12-RAIDT_DSR_Theory_M_v8
• 11-RAIDT_Academic_Logic_M_v11