• workshop_dense_100#slide 12

For the workshop, the priority map should begin with what is most foundational and least negotiable. First comes the problem diagnosis: why model-level documentation, principle lists, and episodic audits leave a gap at the point where configured use actually occurs. Second comes the centre of the framework: the run as the unit of governance. Third comes the run-level evidence pack as the minimum proof object for a material run. Fourth comes the five pillars (Responsibility, Auditability, Interpretability, Dependability, Traceability) and the resulting score profile, because that is where the workshop learns to convert evidence into governance judgement.

What comes later are the layers whose meaning depends on those foundations already being in place. Influence methods as governance interventions come after the score profile, because their value is assessed by the change they produce in governance outcomes and evidence quality. Implementation comes after interventions, because manual, semi-automated, and orchestration-layer designs are different ways of sustaining the same evidentiary discipline. Policy mapping and standards alignment come later still, because they translate the already-defined evidence grammar outward to regulatory and assurance contexts. The empirical programme should usually be discussed last, since it asks how this inner logic performs across domains, repeated runs, and sector calibrations. In short, what comes first is definition and evidencing; what comes later is extension, translation, and cumulative testing. That order keeps the workshop analytical rather than diffuse.

In an HR workshop on shortlist justification, the first discussion should not be whether the organisation is generally 'responsible'. It should be whether one shortlist run can be reconstructed and scored. The team therefore starts with the run record, the evidence fields, and the five-pillar score profile. If Traceability is weak because prompt versions and criteria versions were not retained, the next step is to discuss intervention and implementation choices, such as versioned templates or adapter logging.

Only after that should the workshop turn to broader questions about policy compliance, procurement, or cross-team rollout. By sequencing the meeting in that way, later strategic decisions remain grounded in the evidence architecture of one real use case.

• 11-RAIDT_Academic_Logic_M_v11
• 12-RAIDT_DSR_Theory_M_v8