• workshop_dense_100#slide 56

The E. Pillars & Scoring branch is the part of RAIDT that converts the framework from vocabulary into measurement. Its first task is to define the substance of governance readiness through the five pillars (Responsibility, Auditability, Interpretability, Dependability, Traceability). Its second task is to specify the run-level evidence pack as the scored object, so that judgement is tied to inspectable artefacts rather than to narrative claims. Its third task is to set the anchored scoring logic that reviewers apply to each pillar, with anchors 1=missing / 3=partial / 5=audit-ready. Across the papers, this branch therefore links concept, evidence, and judgement into one operational chain: run as the unit of governance, evidence capture as the proof basis, and structured scoring as the measurement layer.

As a branch overview, it also includes the supporting machinery that makes scoring usable in real organisations. The papers emphasise calibration with worked examples, repeat-run testing for Dependability, partial automation for objective evidence checks, and contextual calibration by sector and workflow. The branch is therefore not a generic scorecard. It is where RAIDT shows how different influence methods as governance interventions change what evidence exists and what score is justified. It also explains why the full score profile is retained even when a composite mean is reported: governance trade-offs matter, and a run that is strong on Auditability may still be weak on Interpretability or Responsibility. In short, the Pillars & Scoring branch is the governance measurement core of RAIDT, because it defines what is being judged, how it is evidenced, how it is scored, and how the resulting profile supports review, escalation, and improvement.

In cybersecurity alert triage, the branch logic becomes concrete. A team can use a structured prompt, approved threat-intel retrieval, and preference constraints to reduce invented indicators. The run-level evidence pack then captures the prompt version, retrieved snapshot IDs, tool outputs, hashes, and repeat-run results. Under the Pillars & Scoring branch, those artefacts are assessed separately: Dependability draws on repeat-run stability, Traceability on source linkage and provenance, and Auditability on whether the run can be replayed and checked.

The result is a profile that tells the team more than whether the output sounded plausible. It shows whether the alert-triage run was governable, where the weak point lies, and which control should be strengthened next.

• 08-RAIDT_Foundations_M_V50
• 11-RAIDT_Academic_Logic_M_v11