Why Decision Velocity Is the Only Metric That Matters in the AI Era

Key points

• MIT research shows 95% of generative AI initiatives fail to achieve their objectives, and RateGain's 2026 State of Distribution report shows only 10% of hospitality AI efforts deliver intended value.
• Bzdawka defines Decision Velocity through three interdependent variables: time to answer, confidence in the answer, and distance to action; all three must function together.
• A hotel is a collection of distinct businesses—lodging, F&B, spas, event spaces and retail—often stacking 10 to 20 specialized systems, each built for a specific operational need rather than real-time enterprise data sharing.
• Hospitality technology remains heavily reliant on legacy batch processing: transactional data is shipped above-property, computed overnight and pushed back down, so AI deployed on this architecture only spotlights the underlying chaos.
• He advocates decoupling the data layer from core transactional systems, extracting raw data via APIs and normalized reporting pipelines into unified cloud environments—a 'lakehouse as a service' model for mid-sized chains.
• VCC settlement case: one major hotel chain suffered a 30% VCC transaction failure rate from manual delays; after deploying an automated pipeline for real-time VCC validation and posting, the failure rate fell to under 1%.
• That workflow shift also eliminated back-office administrative overhead, repaired relationships with OTA partners and spared front-desk staff from uncomfortable guest confrontations.