The New Question of Artificial Intelligence in Hotels: How Much Automation Is Not Enough, and How Much Revenue?

In the U.S. hotel sector, rising prices alone are not enough to maintain profitability. According to Mews’ 2026 U.S. hotel performance data, the average daily rate (ADR) increased by 5% throughout 2025, while RevPAR declined by 5.5%. As 2026 began, performance started to recover, but cost pressures continue. According to data shared by Mews managers with Hotel Management, the average occupancy rate in the U.S. is around 60.5% and has increased by approximately 2.3% year-over-year. The same assessment indicates that while hotel expenses have risen by 5-10% annually, RevPAR has remained flat or grown by about 3%. This gap shows that technology investment is no longer just about modernizing operations; it has become directly linked to profitability issues.

Mews’ data particularly highlights distribution channels. According to the company’s analysis of U.S. hotel performance, OTAs account for approximately 65% of booking volume, while direct bookings generate an average 20 dollars higher ADR and guests who book directly stay an average of 0.3 nights longer. In contrast, offline direct channels such as phone, email, and walk-ins still account for about 25% of all bookings. Since a significant portion of these transactions is conducted manually, opportunities for automated upselling and data collection can be limited.

A similar gap is seen on the upselling side. Mews data shows that the average value of additional sales made at the front desk is 22 dollars, while it rises to 52 dollars during kiosk check-in and 57 dollars in the online booking engine. Nevertheless, only 11% of guests in the U.S. check in online before arrival, and 8% use kiosks. While these data may not yield the same results across all hotels, they indicate that digital offers made at the right moment have the potential to create additional revenue, not just save on staffing costs.

The adoption rate of artificial intelligence is now quite high. According to a global survey conducted by Mews in over 500 hotels from December 2025 to March 2026, 98% of participants reported using AI in at least one part of their operations in the last six months. AI is used in an average of 11 out of 19 common hotel tasks tracked in the study, taking on about 56% of the workload in those tasks. The most common areas include content creation, price optimization, and translation, while broader automation opportunities are noted in some back-office processes such as housekeeping control, workforce planning, and purchasing.

However, widespread use does not mean full automation. In the study, 83% of hoteliers expressed confidence that AI-supported tools help in decision-making, while 59% believe that the greeting and check-in experience at the front desk should continue to be conducted by humans. Mews also reports that 41% of hotels do not yet have an official AI policy. Therefore, the issue facing the industry is not just about purchasing technology; it is about which decisions will be left to automation, who will control the outputs, and how employees will work with these systems.

One of the most significant barriers to technology is the fragmented data structure. The joint emphasis of Mews and Hotel Management is that if the data held in PMS, RMS, CRM, housekeeping, or guest communication systems is disconnected, the potential of AI can be severely limited. If there are incomplete guest profiles, manual Excel files, and systems that do not communicate with each other, AI may work faster, but better decision-making is not guaranteed. For this reason, Mews sees clean data, open APIs, and connected systems as prerequisites for AI investment for hotels.

On the employee side, research does not present a simple picture of “AI will eliminate staff.” Hoteliers’ approach is more about leaving routine tasks to technology while preserving the employee's role in services that require high contact with guests. Rate management, data analysis, translation, and content creation are seen as more suitable for automation, while hybrid models are emerging in concierge, guest communication, loyalty, and some upselling processes.

It should be noted that this data is published by Mews. Since Mews is also a hospitality technology company selling PMS, RMS, payment, and operations technologies, the technology benefits in the research should be considered not as an independent industry average but as findings published by the company. Nevertheless, the assessment made by Hotel Management with industry professionals supports the view that technology should be measured in terms of ROI under cost pressure.

In our opinion: The significant change for hotels in 2026 may not be the widespread use of AI but rather the more rigorous measurement of technology investment outcomes. The simultaneous purchase of chatbot, RMS, automated check-in, housekeeping software, and various analytical tools does not mean digital transformation. First, the hotel’s most expensive or slowest friction point should be identified; then, technology investment should be tracked with measurable indicators such as labor time, additional sales, direct booking share, RevPAR, error rate, or guest satisfaction. In Turkey, especially in an environment where personnel costs, energy, and financing expenses are rising, the value of AI should be measured by how much time the existing team can allocate to higher-value tasks and how much it improves the revenue/cost balance of the business.