I've been running a pre-publish citation check on B2B SaaS content for the last month. The finding: most posts are invisible to LLMs not because they're wrong, but because they're not structured to be extracted.
This post has 5 copy-paste prompts that reveal the gap before you hit publish. Prompt 3 (Comparison Retrievability Test against Claude) has the most dramatic before/after I've seen. Posts that hedge comparisons never get cited. Posts that take a position with a number do. The table in that section is worth reading even if you skip the rest.
The 5 prompts:
- Claim Specificity Test -- Ask Claude if it would cite specific claims from your intro. If it says "no," you'll see exactly why.
- Structure Extraction Test -- Paste your H2s. Ask ChatGPT to summarize your post using only headers. Vague headers = zero extraction.
- Comparison Retrievability Test -- Ask Claude to compare your stance to competitors. Hedged comparisons don't survive the filter.
- Attribution Anchor Test -- Does your post have a named, citable phrase? "X% of operators..." beats "many operators..."
- Freshness Signal Test -- Ask Perplexity if your content reads as current-year or evergreen. Staleness kills citation rate.
Read the full breakdown with copy-paste prompts at: operatoriq.io/blog/llm-citation-test-prompts-before-publish/
