Only 7% of Hotel PMS Vendors Publish Self-Serve API Docs, New Study Finds

A study by the AI Hospitality Alliance and HotelLogic, covering data as of September 1, 2026, audited 343 hotel property management system vendors and found that only 7% — 24 of 343 — publish fully self-serve API documentation. The other 93% keep their APIs gated or entirely undocumented.

Among 18 vendors the study scored in full across ten fixed capability dimensions, composite scores ranged from 80.0 for Mews down to 26.3 for Oracle OPERA 5, with the top five — Mews, Apaleo, Oracle OPERA Cloud, Beds24, and Cloudbeds — all scoring 73.3 or above. The methodology is described as “anchored to publicly available evidence against ten fixed dimensions,” with the study stating sponsor placement “never influences findings, scoring, or vendor placement.” The findings that matter most for AI adoption specifically: “webhook maturity and AI-agent suitability are not yet standard,” and “AI Readiness does not reach the top of the scale for any assessed vendor.” Governance scored even lower — only two of the 18 scored profiles exceeded a 3 on Governance & Security, largely for missing GDPR data processing agreements.

That gap lines up with what a HITEC 2026 recap predicted a month earlier: the next PMS category shift would be defined by governance, not novelty. That piece described an emerging “Agent Management Platform” category built around which models agents use, what data they can access, and audit trails, plus an “AI-native PMS” evolving from a record-keeping system into an operating-system layer. The API study’s finding that no vendor scores well on AI readiness confirms that prediction wasn’t premature — the infrastructure hasn’t caught up yet.

For hospitality technology buyers, the study quantifies a bottleneck sitting underneath every “AI-powered” PMS pitch: even where a vendor markets agentic capability, the API layer that would let a third-party AI agent actually connect to PMS data remains gated or undocumented across the large majority of the vendor landscape, and governance scoring lags capability scoring almost everywhere it was measured. Evaluating a PMS on AI readiness now means asking for the API documentation directly, not the roadmap slide.