We’ve published a new paper arguing that Generative Engine Optimization (GEO) should now be treated as a legacy framework.
GEO was useful between 2020–2024 when AI assistants leaned on structured data, UGC, and citations. But in 2025, it fails for three reasons:
Volatility: 40–60% of AI answers about brands change monthly.
Fragmentation: ChatGPT, Gemini, Claude, and Perplexity all use different pipelines.
Hallucination risk: Even “cited” answers can be fabricated or misleading.
We propose AI Visibility Optimization (AIVO) as the successor framework.
AIVO measures citation density, entity authority, and visibility decay across LLMs and connects visibility loss directly to revenue exposure.
Paper link (SSRN): [insert link once live]
Curious what HN thinks: is GEO worth keeping alive as a term, or should the industry move on?