Making a camps directory the source AI answers quote
US Sports Camps runs Nike Sports Camps across dozens of sports. Their buyers had started asking AI assistants the questions they used to type into Google, and the answers were citing everyone else. We rebuilt the tips section around the questions the engines were actually researching.
The situation
When tracking started in April 2026, AI engines were already reading the tips section. 27.5% of tracked answers across ChatGPT, Gemini, Google AI Overviews, Perplexity and Grok cited at least one tips page. The trouble was what those pages were. Much of what got cited was older, shorter tips content, and several questions parents were asking, such as why some camps cost more than others or how to vet an overnight camp, had no dedicated page at all.
The baseline also showed a naming gap. On unbranded questions about coaching and staff quality, 20.9% of answers cited a US Sports Camps page, but only 2.1% named the brand. The engines were using the content as evidence while giving the credit to no one, or to someone else.
The method
Rather than guess at keywords, we mapped what AI systems were already researching about the brand. Google Search Console exposes AI-generated sub-queries as impressions with near-zero clicks. Clustered, those sub-queries are a map of the questions engines are asking on a buyer's behalf and failing to find a good answer for.
- Pull fan-out sub-queries with high impressions and near-zero clicks, then cluster them into topics.
- Baseline how often the brand is named, and which sources get cited instead, across tracked prompts.
- Write against the clusters where citation rate is high but brand naming is low, since those are answers the engines already want and the brand is missing from.
- Publish in waves, then re-baseline each wave to measure days to first tracked citation.
The result
View as table
What the numbers do not say
Camps are seasonal, and these months run through peak summer, so part of the movement may be demand rather than content. The tracked prompt set shrank to about 820 prompts in June and July before returning to about 1,000, and it was not held fixed as waves shipped. The citation timing compares first Google impression, not publish date, and Scrunch does not run every prompt every day, so read anything under two days as "within a day or two." Share of voice on coaching and training prompts was already 73% in April, so it describes a position held, not a gain. And naming still lags citation: on unbranded coaching questions, 35.5% of answers cited a US Sports Camps page in September, but 3.9% named the brand. That gap is the next piece of work.
Scope and attribution
US Sports Camps works with a separate digital agency across the wider site. Soulcraft's engagement covers the tips section: fan-out query cluster analysis, topic selection, article production, and AI visibility measurement. The figures on this page are scoped to that work and are not a claim about sitewide performance. Nike Sports Camps is provided by US Sports Camps.
Want to know what the engines say about you?
We run a baseline before anyone talks about scope, so you see where you stand first.
Get Started