Swissalytics

Case study: Terre des hommes, +42% AI citations

In June 2025, Terre des hommes reached out to us: they rarely appeared in ChatGPT answers about child aid in Switzerland. By October, they were cited in 42 % more relevant queries.

Starting point

The site was technically clean but invisible to an LLM: no JSON-LD Organization, no FAQPage, named entities rarely present in page body text. Starting Swissalytics score: 58/100.

The eight fixes applied

  1. Complete JSON-LD Organization. With logo, founding date, mission, area of action, and contact. Deployed within 48 hours.

  2. FAQPage on key pages. Seven recurring donor questions, structured in JSON-LD and mirrored as visible text on the page.

  3. Rewritten h1s and meta descriptions. With named entities (name, founding year, countries of action) instead of slogans.

"Eight fixes, zero redesign. The site looks identical to a human visitor — it is unrecognisable to an LLM."

Measured outcome

Tracked across 80 test queries (child protection, aid, Switzerland, Lausanne). Before: cited in 19 answers. After: cited in 27. +42 %. Swissalytics score: 58 → 84.

Read next
GuideRetrieved isn't cited: the discovery layer Google can't seeAnalysisYouTube: the #1 signal for AI visibilityDecisionSchema markup barely moves your AI citations