Recorded prompt
“How do AI search engines choose which sources to cite?”
- Product
- Microsoft Copilot
- Surface
- Microsoft Copilot web chat with web search
- Model
- Copilot Auto (model version not exposed)
- Locale
- en-US (English answer; zh-CN interface)
- Observed answer
- Open session ↗
- Observed
- 15 November 2026, 18:30 CST
- Fresh conversation
- Yes
What the answer said
Copilot described a retrieval-and-ranking process before answer generation: understand and rewrite the question, retrieve candidate pages or passages, rank passages rather than whole pages, evaluate relevance, expertise, extractability, corroboration, freshness and verifiable facts, then attach citations to sources that support specific answer statements. It said a top Google position does not guarantee citation, different AI engines often have low source overlap, and consensus across trusted sources can reduce factual risk.
6 recorded sources
What this observation cannot show
Authenticated Microsoft Copilot web-chat run in a fresh conversation. The answer is one observation within a small repeated sample and does not expose Copilot's proprietary ranking weights.