• integrated_82#Q4.20

In RAIDT, interoperability means that governance evidence can be compared and reused across different laws, standards, and organisations without requiring those regimes to become identical. The policy pathways paper is explicit that RAIDT is not a competing legal regime; it is an evidence layer. The audit and accountability paper explains why this matters conceptually: governable generative AI use requires a bounded proof object that can be reconstructed, reviewed, and challenged after the event. RAIDT supplies that object by treating the run as the unit of governance and by attaching to each material run a run-level evidence pack and a score profile.

This matters for policy because regulators and standards bodies increasingly ask organisations to show operational proof rather than broad ethical aspiration. It matters for audit because internal audit, incident review, complaint handling, and quality review all need to reconstruct what happened in a specific use event, not merely confirm that a policy exists. It matters for procurement because buyers and suppliers often do not share tools, access rights, or internal processes, yet they still need a common evidential language for tender requirements, audit rights, and post-incident escalation. RAIDT's five pillars (Responsibility, Auditability, Interpretability, Dependability, Traceability) provide that language. They let a buyer ask for governance-grade run records, an auditor sample runs like transactions, and a policy team align evidence to multiple frameworks. Interoperability is therefore practical rather than rhetorical: it reduces duplication, supports contestability, and makes cross-framework assurance more defensible.

Take a finance adverse-action workflow in which a bank uses generative AI to draft an explanation for a credit-related decision. Procurement needs supplier commitments, audit needs replayable evidence, and policy staff need to show alignment with several governance frameworks. Without interoperability, each function may ask the vendor for different documents and still lack a usable record of one disputed case.

With RAIDT, the bank requires a run-level evidence pack containing the prompt version, model deployment, reason-code inputs, provenance for retrieved material, output hash, reviewer checks, and escalation record. Procurement can turn that structure into contractual evidence obligations, internal audit can sample the run later, and the policy team can map the same evidence to legal and standards expectations. One pack serves three governance functions because the evidential core is shared.

• 10-RAIDT_Policy_Pathways_M_V50
• 16-RAIDT-Audit-Accountability_M_v05