From Invisible to Cited - How AiVIS Cite Ledger Went from 0 Citations to Consistent AI Answer Presence | AiVIS Cite Ledger Blogs

Key Takeaways

• Baseline: Zero AI citations despite strong SERP rankings - verified through citation testing across ChatGPT, Claude, and Perplexity.
• Intervention: Restructured core product pages with FAQPage + HowTo schema, added entity-clear author markup, and deployed BRAG-audited heading hierarchies.
• Result: Consistent AI answer presence tracked through the Cite Ledger with evidence IDs linking each structural change to citation outcomes.
• Key learnings: Schema density, answer-block clarity, and author credibility were the primary levers - each measured through the SSFR evidence framework.

Article

We started where most sites start in the answer era: invisible. AiVIS Cite Ledger ranked respectably in traditional search and got cited by essentially zero major answer engines. Not few. Effectively zero. We knew this because we tested it, asking ChatGPT, Claude, and Perplexity the questions we should have owned and watching them name everyone but us. Then we ran our own methodology on ourselves, tracked every change through the Cite Ledger, and watched what moved.

This is a case study about applying our own audit to our own site, which is a uniquely uncomfortable exercise, because there is nowhere to hide. The tool either explains your invisibility with evidence or it does not. What follows is the honest version: the baseline, the specific structural changes we made, the results we could verify, and the parts that mattered far more than we expected. We are sharing it because the pattern generalizes, and because a case study with no evidence trail is exactly the kind of unverifiable claim our own methodology is built to reject.

The baseline: ranked and uncited

The starting condition was the one that confuses every team that hits it. Our rankings were fine. For several of our core topics we appeared in traditional search results where you would expect a credible source to appear. And yet when we posed the natural-language questions those pages were meant to answer to the major answer engines, our pages were absent from the synthesized responses. The engines answered confidently, drew from other sources, and never reached for us. Ranking presence and citation presence were completely decoupled, exactly as our methodology predicts they can be.

We did not treat this as a vague impression. We ran citation testing across ChatGPT, Claude, and Perplexity, recorded which sources got named for which queries, and established a concrete baseline: the specific questions where we should have been cited and were not. That baseline mattered, because without it any later imp

Enable JavaScript for the full interactive reading experience with related articles and discussion.

Cited external sources