Can Claude find your business without a direct link?
- Published:
- Reading time:
- 2 min read
Use two separate Claude tests to find out whether it can discover a business and read a service page when given the address.
Test Discovery and Page Access Separately
If Claude can explain your service when you paste a link but leaves you out when a buyer asks for options, run two separate tests. The first checks whether it discovers your business from the question alone. The second checks whether it can read the exact page you want buyers to see.
01Start with the buyer’s question
Imagine a company that repairs commercial coffee machines in Porto. In a fresh Claude chat with web search enabled, ask which providers cover Porto and explain whether they offer on-site repairs for cafés. Leave the company name and URL out. Save the answer and open its sources. Note whether it mentions the business, which page it links to and whether the source supports the on-site claim.
02Then give Claude the exact page
Start another fresh chat with web search enabled. Paste the public service page URL and ask Claude to read it and identify where the business works, which machines it repairs and whether it offers on-site service. Ask it to point to the text it used. Check the answer against the page yourself. If Claude cannot open the page, record that separately from a result where it opens the page but cannot find the requested detail.
| Discovery test | Direct page test | Check next |
|---|---|---|
| Business absent | Page opens and answers the question | Inspect the sources shown for the buyer question. Check whether your page is easy to find from your own navigation and other relevant pages. |
| Business absent | Page will not open | Ask your web team to check access and security rules for the exact URL. |
| Business appears via a directory | Page opens but omits service coverage | Put the verified coverage on the page customers should use to enquire. |
| Business and page both appear | Page gives the expected details | Keep the result as a baseline for that question and date. |
Reading a page you supply and finding it during a buyer’s search answer different questions.
03Ask your web team to check the right setting
Anthropic documents separate web agents. Claude-SearchBot supports search discovery. Claude-User retrieves content when someone asks Claude to look it up. ClaudeBot is associated with model training. Before changing robots.txt, a public file that tells automated visitors which pages you want them to crawl, explain which of the two tests failed. A missing recommendation alone does not show that any agent was blocked.
If you can reproduce a gap, share the two prompts, the linked sources and the affected page. Studio 17’s SEO and AI visibility work can help separate a page access problem from a page that leaves a buying question unanswered.
Reading a page you supply and finding it during a buyer’s search answer different questions.
Pedro Leonardo
Find where Claude loses the page
Share the buyer question, the page you supplied and what happened in both tests. We can help identify which checks to run next.