K-means clustering is an unsupervised learning technique that groups data based on similarity. Discovering a cluster with high spending but low product diversity is a plausible and meaningful business insight: it may represent loyal customers who repeatedly purchase a narrow range of products . The auditor should therefore recommend treating this cluster as a potentially valid segment (B), subject to further business analysis and controls where appropriate.
Option A is incorrect because this pattern does not imply algorithm failure. Option C (adding more clusters) might overcomplicate the segmentation without evidence that the current clustering is deficient. Option D misunderstands the purpose of clustering; a supervised model would require labeled outcomes and is not necessarily "more accurate" for exploratory segmentation. AAIA's content on AI in audit processes stresses that auditors must interpret AI-driven insights critically, not assume anomalies equal errors.
References:
ISACA, AAIA Exam Content Outline - Domain 3: AI in Audit Processes (AI tools, use of clustering and analytics in audit).
ISACA analytics guidance on interpreting unsupervised learning outcomes in an audit context.