Hotel Leaders Shift AI Focus from Hype to Targeted Operations

Hotels are deploying artificial intelligence by shifting away from unstructured experiments toward targeted operational fixes and unified guest platforms. While most properties lack a formal plan, industry leaders advise starting with internal administrative tasks, organizing existing reservation data, and using predictive tools to track evolving consumer preferences before introducing automated systems directly to guests.

Why do most hotel AI projects fail to deliver returns?

The vast majority of generative artificial intelligence pilots in hotels yield zero measurable return on investment because they occur without a formal plan. According to research by Prof. Roland Schegg of HES-SO Valais-Wallis and the German Hotel Association, unstructured pilot projects result in up to 95% of generative AI tests producing zero measurable return on investment [1]. Straiv reported that while over 70% of surveyed hoteliers view AI and digital technologies as crucial to their success over the next two years, only 9% have anchored these topics in a formal strategy document [1]. Instead, 35% of properties test AI tools on a case-by-case basis, and 31% simply follow trends [1]. The German Hotel Association data shows that implementing new tools in small and medium-sized properties often fails due to legacy, server-based property management systems that lack modern application programming interfaces [1].

Where should general managers begin their implementation?

General managers should begin by identifying tasks where their teams spend time on work that does not require a human. Writing for HospitalityNet, Matthijs Welle, CEO of Mews, listed repetitive guest messaging, manual reporting, rate updates, and back-office administration as the best starting points [2]. Jacqueline Nunley, Global Industry Strategist at Salesforce, advised that hotels should start internally rather than in front of guests [2]. She suggested using AI within existing tools to summarize shift handovers, surface updates, and answer operational questions [2]. Gustav Søgård, Founder of Opally, recommended picking a high-volume workflow like pre-arrival emails or common website questions, measuring response times before beginning, and running a focused pilot [2].

How does data quality impact the success of automated tools?

Artificial intelligence systems require accurate, property-specific information to function correctly. Welle stated that an AI system that does not know a hotel's rate structure, guest mix, or workflow will produce answers that sound right but are factually incorrect [2]. Søgård noted that AI requires access to a property's policies, room details, tone of voice, and live rates from the property management system or booking engine [2]. David Turnbull, Founding Partner at ATworld, argued that hotels should not wait for perfect data, but instead use the best available information for a tightly defined use case and improve it over time [2]. Nunley pointed out that if a returning guest is still asked if it is their first time staying, applying AI will only scale that existing problem [2].

What is the "Amazon cart" approach to guest journeys?

The "Amazon cart" approach combines guest preferences, ancillary bookings, and artificial intelligence into a single platform where travelers can select multiple activities and complete one checkout. Forbes reported that Shawn Tarter, CEO of RealTime Reservation, is building this model to replace the fragmented system where spas, restaurants, and children's programs operate as separate businesses [3]. RealTime Reservation recently combined with hospitality guest-experience platform STAY, expanding the company's reach to approximately 2,500 hotels globally [3]. Tarter explained that allowing guests to plan their entire trip in one place before arrival gives hotels more time to manage inventory and coordinate outside providers [3]. This model extends yield management beyond room rates to include restaurants, cabanas, and spa treatments [3].

How are design firms predicting future guest preferences?

Hotel design firms are partnering with artificial intelligence startups to test concepts and predict what consumers will want next. CoStar reported that Ron Swidler, CEO of the Gettys Group, uses an AI startup called Vurvey Labs as a predictive tool [4]. Vurvey Labs builds databases combining real responses and video responses from consumers testing products for brands like Unilever, Procter & Gamble, Dove, Nike, and Legos [4]. Swidler explained that the system analyzes these responses to determine what is most important to consumers now and what will be most important to them later [4].