People aren't Googling anymore. They're asking AI.
What you get
12,000 prompts
Real buyer prompts across 40 verticals - legal, medical, home services, SaaS, retail, finance.
6 major models
ChatGPT, Gemini, Claude, Perplexity, Copilot, and Apple Intelligence sampled and scored.
Per-model breakdown
Citation source diversity, freshness bias, and authority thresholds vary massively by model.
Quarterly updates
Recomputed every quarter as models ship. What ChatGPT cited in Q1 isn't what it cites now.
Which sources does each model favor?
Aggregated citation preference from Q1 2026 sampling. Rank order approximate - actual composition shifts weekly.
ChatGPT
Wikipedia, established publishers, brand's own domain, industry associations
Perplexity
Fresh news publishers, Reddit, YouTube transcripts, Substack, arXiv
Gemini
Google Knowledge Graph, Google Business Profile, YouTube, high-authority publishers
Claude
Long-form journalism, academic sources, established brand documentation
Copilot
Bing-indexed publishers, LinkedIn, Microsoft-network sources
Perplexity Shopping
Reddit, YouTube reviews, specialty publications, brand-owned product pages
Methodology summary
- • 12,000 real buyer prompts drawn from client engagements and public search-intent datasets.
- • Each prompt run against 6 major AI models via public interfaces (no scraping of gated APIs).
- • Citation sources parsed from linked-source blocks and inline attributions.
- • Deduplication by domain; sub-paths counted separately when the model surfaced them independently.
- • Recomputed quarterly. Full methodology available on request.
Ready to get recommended by AI?
Allocation is capped. Apply now to see if your business qualifies.