TwistLens is a docent-informed, AI-supported image transformation system that creates spoiler-minimizing previews for pre-visit museum experiences. The method extracts informational cues from docent text through a structured taxonomy and applies transformation strategies (EchoLens and DecoyLens) to balance anticipation, interpretation, and representational coherence.