SiteMinder's New AI Engine Keeps a Human in the Loop On Purpose
SiteMinder's Dynamic Commerce Engine surfaces prioritized pricing and distribution recommendations across its 56,000-property network but requires operator approval before executing — the same human-in-the-loop design showing up in Duetto's 95% ML forecast accuracy and Mews' 13% Autopilot revenue lift.
SiteMinder says the execution gap in hotel distribution isn’t a data problem anymore — it’s that hoteliers get reports, not decisions. Its new Dynamic Commerce Engine uses machine learning to continuously scan pricing and distribution data and surface prioritized recommendations instead of raw dashboards, while also flagging distribution-layer problems like broken channel connections or unmapped room types. The system is built with a human explicitly in the loop: recommended changes require operator approval before execution, positioned as augmentation rather than full automation. SiteMinder is running the engine across its existing scale — roughly 56,000 properties and 2.6 million rooms, processing close to 4 billion availability-rates-inventory signals alongside 140 million reservations and 300 million room nights a year. The announcement cites survey data showing 51.4% of travelers abandon a hotel booking due to a poor online experience, and 26.5% cite distrust in reviews, photos, or listing details. CEO Sankar Narayan: “Reports and AI assistants are no longer enough. Hotels need to know which opportunities actually matter.”
That human-approval design isn’t unique to SiteMinder — it’s where the category is converging. Duetto’s new forecasting engine hit 95.22% average forecast accuracy across its customer portfolio, a 12% improvement in sMAPE, and added a “driver waterfall chart” specifically so revenue managers can see which factors are driving a number rather than trust a black box. And a 6,000-hotel study of Mews’ Autopilot pricing feature found a 13% lift in revenue per square meter over 18 months — but also that only 55% of eligible Mews customers currently run it in full-autopilot mode, with the rest still keeping a hand on the wheel.
Read together, the pattern across 2026’s revenue and distribution AI is consistent: vendors are optimizing for prescriptive, explainable recommendations that hoteliers approve, not fully autonomous pricing or channel decisions — because the buyers themselves, per the Mews adoption data, aren’t ready to hand over full control yet.
Source: Hospitality Net Auto-generated brief — verified before publishing.