Real example, anonymized: name, website address and identifying details removed. Findings are from the original report of Nov 2, 2026.

City hotel in northern Italy: agent ready website reading

Passive analysis of public pages, 10 pages read

Score: 3/10

Why it matters: direct bookings

Goal: more direct bookings, fewer commissions to booking portals. More guests ask Google, ChatGPT or Perplexity where to stay. When the website gives clear answers, the assistant sends the guest straight to the hotel; when data is missing, it sends them to a portal that takes a commission on every night.

What works well

• fast homepage: 1.35 seconds on average
• practical information (cancellation, check in, parking, pets) easy to find
• robots.txt allows the main AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot)
• clear booking path through the hotel's booking engine

Priority fixes

1. Hotel structured data. The homepage does not declare the property as a Hotel in structured data (types found: none). Add a JSON-LD block of type Hotel with name, address, phone, email and languages.

2. Contacts in structured data. Address, phone and email are not given together in structured data, so an AI assistant has to guess them from the text.

3. llms.txt file. The site has no llms.txt: a short text summary of the property (rooms, services, contacts, key links) written for AI assistants.

4. A perk for booking direct. For example a welcome drink for guests who book on the website instead of a portal, stated clearly on the site: AI assistants look for concrete reasons to recommend a property.

Note: 121 of 122 images have no alt text.

Method: robots.txt respecting crawler, one second between requests, at most 80 pages, speed as the median of 3 loads. TechAlpino, a service of Adige Consulting LLC, info@techalpino.com

Italiano · English · Deutsch · Free audit