At Skift Data + AI yesterday, Seth Borko said on the stage that full AI booking may be three years out—but AI search & discovery is here today. Our logs show AI bots hitting hotel websites every day. Here’s what you need to do now:
If you see Google-Extended, meta-externalagent, GPTBot & ClaudeBot on your web logs, they’re doing two things. They’re training, and they’re mapping. If those bots can’t read about your hotel and learn they can ask for the information travelers want, they’ll steer guests to an OTA site or another hotel website that is readable.
Based on our research, emerging best practices, and the information we’re sharing among Kismet’s peer companies, websites that win in this era will require:
1️⃣ Prices & availability ― Your CRS already has this info, but many hotels don’t have it on their own website. They link out to another booking engine site. This is a problem. Expose this info through a gated /ask or MCP tool so trusted bots learn your price point (and soon quote it).
2️⃣ Need-to-know details ― Menus, spa hours, floor plans, one sheets. We need to get these PDFs and into a searchable store so answers stay current and the PDFs can be delivered when the agents ask. They’re going to pass them to users.
3️⃣ Credible social proof ― Recent reviews and real guest photos via social and reviews sites—good and bad—an emerging consensus is that AI will ask you to provide social proof and the context around it.
Quick homework: take 30 minutes with marketing and revenue to list (a) where your live prices sit, (b) which PDFs your sales, marketing, F&Bs, and events teams use, and (c) which review sites and social feeds impact guest discovery. That sheet is the start of your roadmap to NLWeb readiness.
If you want our full checklist we’re using with hotels and web development teams, you can find a link to download it in the “Visit my website” button in my profile.
See you again next week!