Evidence Engine
Platform Evidence
Each AI platform has different retrieval engines, biases, and citation behaviors. Here's the evidence mapped to each.
Platform-specific optimization
A tactic that works for Perplexity might not move the needle for ChatGPT. We track evidence separately by platform to ensure recommendations are targeted.
ChatGPT
Bing Integration
Evidence
>32,000 referring domains for 2× citation frequencyBOX 3
Domain authority strongly correlates with citationBOX 3
Wikipedia mentions boost corroboration signalsBOX 3
Tactics
→Build domain-level authority through backlinks
→Get mentioned in Wikipedia and authoritative directories
→Focus on high-DA content partnerships
→Prioritize pages that already rank well in Bing
Perplexity
Custom Index + Reddit Heavy
Evidence
46.7% of all citations come from RedditBOX 3
55% Top-3 rate with Q&A headingsBOX 3
Strong recency bias in source selectionBOX 3
Tactics
→Establish genuine Reddit presence in relevant subreddits
→Structure content with H2s as direct questions
→Update content frequently — freshness matters more
→Include explicit Q&A format sections
Claude
Brave Search
Evidence
1.7× citation lift with limitations/caveats sectionsBOX 3
Constitutional AI rewards intellectual honestyBOX 4
Nuanced analysis preferred over absolute claimsBOX 3
Tactics
→Include explicit limitations sections
→Present balanced perspectives
→Add nuanced analysis rather than absolutes
→Show uncertainty where appropriate
Cross-platform factors
Some evidence applies across all platforms:
Schema markup81% correlation (Princeton GEO) — universal
Content freshness70% of cited pages change within 2-3 months
Page speed3.2× lift at FCP < 0.4s — affects crawl success
Answer nuggets+40% visibility with extractable facts