AI 感知指数 2026：大型语言模型如何在 AI 时代定位品牌

Key Points

• The aim is to develop a comprehensive framework to measure brand perception through large language models.
• Introduced AI Perception Index (MPI) as a scoring framework.
• Benchmarked 15 brands across five sectors: AI tools, CRM platforms, marketing automation, analytics, and e-commerce.
• Applied structured query protocols to GPT-4o and Claude Sonnet 3.7.
• Identified significant cross-model perception drift.
• Highlighted disconnect between surface mention frequency and semantic authority.
• Established Perception Control as a new strategic discipline beyond Generative Engine Optimization.

Abstract