• workshop_dense_100#slide 2

The 100-slide workshop deck should be used as a guided supervisory route through the RAIDT architecture rather than as a linear recital of all material. A sensible use pattern is to begin by establishing the problem statement: why policies, model cards, and episodic audits are insufficient when governance failures arise in one configured use. The deck should then bring the group quickly to the centre of the framework: the run as the unit of governance and the run-level evidence pack as the minimum proof object for material uses. This ensures that discussion begins with reconstructability, not with general attitudes about responsible AI.

After that, the deck is best navigated ring by ring. First, the group should inspect how the run-level evidence pack supports the five pillars (Responsibility, Auditability, Interpretability, Dependability, Traceability) and produces a score profile using anchors 1=missing / 3=partial / 5=audit-ready. Second, it should examine how structured prompting, retrieval augmentation, PEFT/LoRA, preference-based alignment, and stacked configurations operate as influence methods as governance interventions. Third, it should turn to implementation choices: manual pilot, semi-automated logging, or orchestration-layer capture. Only once those inner and middle layers are understood should the workshop move to policy alignment, sector calibration, and the empirical programme. Used this way, a large deck becomes a supervisory decision aid: it tells readers what must be understood first, what can be deferred, and what follow-on reading depends on earlier agreement.

In a public-service workshop on eligibility advice, the facilitator could use the deck in three passes. The first pass asks one simple question: can the team reconstruct one disputed advice run? The second pass checks the score profile across the five pillars, especially Auditability and Traceability, by inspecting whether the exact rule text, retrieval snapshot, model version, and review notes are present. The third pass asks what intervention or implementation change is needed.

That approach prevents the workshop from becoming a policy seminar detached from operational evidence. If the team discovers that retrieval was used but snapshots were not stored, the immediate next slides are the intervention and implementation slides, not the outer policy material. The deck is therefore used diagnostically, not performatively.

• 11-RAIDT_Academic_Logic_M_v11
• 12-RAIDT_DSR_Theory_M_v8