• workshop_dense_100#slide 13

The 'A. Core & Scope' branch should be understood as the entry branch that stabilises what RAIDT is before any later branch explains how it works. In the wording papers, this branch contains the core definition and purpose, context-specific meaning and use, and scope and unit of analysis. It establishes that RAIDT is a governance method for organisational GenAI use, that it produces a run-level evidence pack and a score profile, and that it is intended for settings where outputs can influence decisions, records, communications, or services. This branch therefore defines the framework's object, purpose, and applicability.

Its second function is boundary-setting. 'Core & Scope' explains why run as the unit of governance matters, what counts as a run, and which stakeholders may later need to review it. It also clarifies that RAIDT is strongest where systems are used repeatedly, across teams, or in high-impact contexts, and that it does not replace domain safety practice or formal sector duties. As a branch overview, then, 'A. Core & Scope' gives the minimum conceptual architecture required for every other branch to remain coherent. Without it, the five pillars can be mistaken for generic values, the run-level evidence pack for a logging format, and scoring for a detached compliance exercise. With it, later branches can be read as coordinated parts of one run-level governance framework.

A healthcare organisation planning a note-summarisation pilot would begin with the 'A. Core & Scope' branch before building templates or dashboards. The branch would clarify that the framework applies because the output could influence clinical records and subsequent decisions, and that the relevant governed entity is each configured use, not the model in general.

From there, the organisation can define what a run includes in its setting: the prompt, model configuration, any retrieval, the summary output, and the human check. It can also identify the review stakeholders, such as clinicians, auditors, and governance staff. That early scoping work reduces later confusion because every evidence and scoring decision is anchored to a clearly defined RAIDT object and use boundary.

• 00-RAIDT_Wording_v2
• 11-RAIDT_Academic_Logic_M_v11