A non-coding product lead must choose a Copilot model to explain a legacy billing workflow from a large repository to finance stakeholders, prioritizing clear rationale and traceable explanations over raw coding speed. Which model selection consideration best guides an explainable choice?

The scenario centers on a non-coding lead communicating legacy billing logic to finance, where success depends on plain-language rationale and traceability rather than throughput. Reasoning-optimized models articulate intermediate steps, cite sources, and qualify uncertainty, supporting auditability and stakeholder trust even at the cost of latency, which directly matches the stated priority. A tuning centered on rapid repository-wide navigation aids engineering discovery but optimizes for locating patterns quickly rather than translating causality into accessible explanations, leaving the finance communication need unresolved. Broad context handling that retains extensive history for concise summaries addresses scale and tone, yet retention plus brevity alone does not ensure step-by-step justification or source linkage, risking plausible but unverifiable overviews of intricate billing rules. Specialization in executable prototypes demonstrates behavior through runnable artifacts, which helps developers but substitutes demonstration for explanation and assumes technical fluency finance colleagues do not need. As an exam caveat, distinguish capability from suitability: speed, context size, and generation quality are strengths only when aligned to the explicit criterion, and secondary benefits do not override a primary explainability requirement. As a practical check, pilot candidates on one billing edge case and have finance reviewers reconstruct the why and provenance without help.