Field report · carl.golf
From under 5 to 200+ AI impressions a day, in 8 weeks
carl.golf, a Boston-area golf tee-time booking site, went from nearly invisible in AI search to 200+ impressions a day and more than ten organic customers in eight weeks, on the back of two fixes from a Ron audit.
<5 → 200+
impressions / day
10+
organic customers
8 weeks
one report, two fixes
The starting point
Carl had a real product and a clean site, and almost no presence in AI answers. When buyers asked ChatGPT or Gemini where to book a tee time near Boston, Carl did not come up. The category went to the big aggregators every time. The Ron audit scored Carl low and, more usefully, said why.
Fix one: stop blocking the crawlers
The audit found that robots.txt was blocking AI crawlers. The engines literally could not read the site, so they had nothing to say about it. This is the most common finding Ron sees, and it is a one-line change. We unblocked the crawlers and the door was open.
User-agent: GPTBot Allow: / User-agent: Google-Extended Allow: /
Fix two: write a real llms.txt
Opening the door is not enough if there is nothing clear to read. We added an llms.txt that said, in plain language, what Carl is, who it serves, and why someone would pick it. That gave the engines an accurate, quotable description instead of leaving them to guess from scattered page copy. Mentions followed.
The chart has been hockey-sticking ever since. Two changes, both shipped in an afternoon, and the engines finally started recommending the thing they had been ignoring.
Why it compounded
AI visibility builds on itself. Once the engines could read Carl and had a clear description to repeat, each new answer that mentioned Carl made the next one more likely. The two fixes were small. The compounding was not.
You can read the live audit for Carl any time. See the full carl.golf report →
Questions
What were the two fixes?
First, one line in robots.txt that was quietly blocking AI crawlers, so the engines could finally read the site. Second, an llms.txt that explained in plain language what Carl does, who it is for, and why, so the engines had something accurate to repeat.
How long did it take to see results?
The chart started moving within weeks and compounded over roughly eight weeks to 200+ impressions a day and more than ten organic customers.
Can I expect the same numbers?
No two sites are identical, and Carl started from a low base with an easy blocker to remove, which is the best case. But blocked crawlers and missing llms.txt files are extremely common, so the type of fix transfers even when the exact numbers do not.
How do I find my version of these fixes?
Run a Ron audit. It checks your robots.txt and llms.txt, runs 36 grounded queries across ChatGPT, Claude, and Gemini, and hands you the ranked fixes for your specific site.
Find your two fixes.
~5 min scan · $39 · refunds if useless
Run my audit →