AI tools give different answers to the same question. A new SparkToro study shows this happens often.
The research tests ChatGPT, Claude, Gemini, and Perplexity. Results change nearly every query.
This creates big challenges for brands and marketers. The findings come out in January 2026.
AI Recommendations: Why AI Answers Keep Changing
SparkToro runs the same prompts many times. Each model answers differently almost every run. This makes it hard to trust AI for consistent recommendations.
- ChatGPT changes answers in 70–90% of repeated queries
- Claude varies 65–85% of the time
- Gemini shifts in 60–80% of tests
- Perplexity shows changes in 55–75% of runs
- Same prompt gets different brands ranked first
The study uses commercial questions like “best CRM for small business”. Top results flip often. This surprises marketers who want stable visibility.
Also read about: AI Market Size Statistics 2026: The AI Revolution
What This Means for Brands and SEO
Brands struggle to appear reliably in AI answers. Old SEO rules no longer work alone. Marketers must adapt fast.
- Focus on broad authority across many sources
- Build presence on Reddit, YouTube, and forums
- Create deep, helpful content AI can cite
- Track AI responses to key prompts weekly
- Use multiple models to spot patterns
SparkToro says consistency matters less now. Influence across the web grows more important. Brands that show up often win. The study warns: AI discovery changes the game. Prepare for frequent shifts in 2026.
More News To Read: