31 answers served in 16 days.
This is a UK address-tracing service we launched for solicitors and individuals. It went live on 29 September 2026 as a brand-new domain: no history, no backlinks, no paid links. By the morning of 14 October its pages had been fetched to answer 31 questions: ChatGPT 28, DuckDuckGo AI 2, Perplexity 1. Source: the server logs, read on 14 October 2026.
Quotable89% of ChatGPT's answers came after its search crawler took 35 of the pages, and answers a day rose 10x. The curve turned the day after the crawl.
5 on 10 October, 5 on 11 October, 9 on 12 October, 4 on 13 October and 2 by 08:30 UTC on 14 October. Before that: 2 on launch day, 2 on 3 October, 1 on 5 October, 1 on 9 October.
- 29 September
The site goes live, indexable, with server-rendered pages, dated guides, llms.txt and a sitemap. Two answers the same day.
- 18 September to 10 October
ChatGPT's search crawler (OAI-SearchBot) fetches 35 pages across seven active days; OpenAI's main crawler fetches 52.
- 10 to 14 October
Answers served a day rise 10x. On 14 October the first visitors arrive from ChatGPT: two to the home page, one straight to the instruction form.
For fifteen days the number was zero, and we showed the zero. ChatGPT answers in place; what the person takes away is the name. On 14 October the first visitors clicked through from ChatGPT: two to the home page, one straight to the instruction form. That is why answers served is the reading, and visits from assistants the smaller line beneath it.
Pages ChatGPT read, and what the person most likely asked.
| Page | Answers | What the person most likely asked |
|---|---|---|
| / | 21 | The brand, or the category: "address tracing service UK", "tracing agent UK" |
| /pricing | 2 | What a tracing agent costs |
| /use-cases/trace-a-debtor | 1 | Finding a debtor's address |
| /use-cases/reconnect-with-someone + /about | 1 | Finding a lost person, and who is behind the service |
| /for-solicitors | 1 | Tracing for a law firm |
| /developers | 1 | A tracing API |
| /how-it-works | 1 | How an address trace works |
DuckDuckGo AI cited /guides/finding-missing-beneficiaries and /how-it-works; Perplexity cited the home page. One ChatGPT answer on 11 October read /about and /use-cases/reconnect-with-someone together. Questions are inferred from the page fetched and marked likely; the question text itself is only visible in Bing Webmaster and Search Console AI reports.
| Index crawler | Sitemap pages read | Share of the site |
|---|---|---|
| Amazon Alexa | 125 of 125 | 100% |
| Meta AI | 121 of 125 | 97% |
| Bing | 105 of 125 | 84% |
| Apple | 86 of 125 | 69% |
| 54 of 125 | 43% | |
| OpenAI | 44 of 125 | 35% |
| ChatGPT search | 28 of 125 | 22% |
| Anthropic | 21 of 125 | 17% |
| Perplexity | 13 of 125 | 10% |
| ByteDance | 9 of 125 | 7% |
A fetch shows a crawl, not index membership. Bing feeds Copilot, ChatGPT search and DuckDuckGo; Google feeds AI Overviews and Gemini.
Search Console, 28 days to 2026-10-08: 8 clicks, 543 impressions, average position 34, 38 impressions inside Google's AI features. Real visitors a day: 60 to 110. Backlinks from other sites: 0.
Google treats a new domain as new. The assistants read it the week it went live.
Everything the Audit writes for a client, applied from day one: the site was built the LLMs Bureau way before it had a single visitor. The method is not published. Its readings are on this page.