In mid-October 2025, ChatGPT rolled out a significant model update that materially changed how brands appear in AI-generated answers. Using Answer Engine Insights, we analyzed millions of prompts across ChatGPT and other leading answer engines to understand what changed and why.
What changed
- Brand mention frequency per response decreased by an average of 34%
- Responses shifted from multi-brand lists toward single-recommendation answers in many categories
- The citation pool narrowed — a smaller set of authoritative sources now captures a larger share of citations
- Comparative content (brand vs. brand) saw the most significant drop in multi-entity mentions
Category-specific impacts
The impact was not uniform. Finance, healthcare, and legal categories saw the most significant reduction in brand mentions, likely reflecting increased caution around high-stakes recommendations. Consumer product categories — electronics, apparel, home goods — were less affected and in some cases saw increases in specific brand recommendations.
What to do now
The update reinforces a trend we've been tracking all year: AI models are becoming more selective, not less. Brands that have built genuine topical authority — deep, accurate, well-structured content — are holding their position. Brands relying on volume-based content strategies are seeing the most significant declines.
If your category is one of the affected ones, now is the time to invest in content quality over content quantity.