Half of Hotels Use AI. Fewer Than 1 in 10 See Real Impact From It.

Roughly half of hotels now use some form of AI, but under 10% report seeing real business impact from it, according to the State of Distribution 2026 study covering 343 PMS vendors, more than 270 hotel brands, and over 58,000 properties. The gap traces to three causes: fragmented data architecture that keeps AI tools from seeing consistent, unified guest information; rollouts launched without success metrics defined up front; and vendor ROI claims hotels never independently verify. A separate integration-friction finding sharpens the diagnosis — only 7% of PMS companies publish fully public API documentation, with 93% keeping closed or partially closed APIs, which the study ties directly to why deployed AI tools underperform once installed.

The gap isn’t just technical. Lighthouse’s own response, launching embedded “Ernest Crews” teams that spend 30 to 60 days inside a hotel’s commercial department deploying its Ernest AI teammate, cites its own version of the same divide: only 10% of hospitality leaders feel fully prepared for AI even though 67% expect it to have a major business impact. Lighthouse’s bet is that the bottleneck is organizational, not technological — its crews build custom connectors and automations tailored to a property’s actual workflows rather than shipping self-serve software and leaving hotels to figure out integration themselves.

There’s a second shift buried in the same study worth hoteliers’ attention: AI systems embedded in booking and search platforms now read guest reviews directly to answer traveler queries, meaning a detailed, specific review now performs better in AI-powered search results than a generic five-star rating. Hotels optimizing review responses for a five-star average, rather than for the substance a machine will parse and surface, are optimizing for the wrong reader.