Blab Tech
AIs Dual Impact on Hospitality Unfolds at HITEC 2026
The 2026 HITEC conference in San Antonio highlighted how artificial intelligence is reshaping hotel distribution and operations, while raising questions about cost, brand control, and staffing. The event’s opening days were marked by a mix of optimism and caution.
In the “Distribution 2.0” session, Fredrik Sjoberg of Maison, an AI‑concierge platform, argued that by 2030 most hotel bookings will be made by autonomous agents rather than human travel agents. He noted that human agents historically captured about 10 % of booking revenue, OTAs 15–25 %, and app stores 30 %. Sjoberg cautioned that distribution costs are unlikely to fall and urged hotels to make content machine‑readable, including detailed pet policies, and to seek machine‑connectable endpoints.
The afternoon panel, “The Dark Side of AI,” focused on the environmental and cognitive costs of rapid AI adoption. Panelists cited a 2025 IEA report that AI data‑center electricity rose 17 % year‑on‑year, with most power still coming from coal and gas. Carl Winston of San Diego State, who certifies sustainable resorts, said that AI’s energy footprint complicates sustainability claims. The panel also referenced a study showing that users who let AI write content had poorer recall of the material.
Brand survival was debated in a headliner session. Lennert de Jong of Another Star and Scott Strickland of Wyndham argued that loyalty programs remain effective because AI can become a booking channel that still rewards points. Floor Bleeker countered that if agents talk to agents, hotels may lose distribution control and must compete on guest experience instead. The discussion also touched on whether hotels need staff at all; Bleeker suggested staff can be a barrier to service, while Keryn McNamara of Aimbridge insisted that hospitality is fundamentally a human business.
On day two, James Taylor emphasized that AI should augment rather than replace human curiosity. He cited research indicating that fear of appearing less capable deters use, especially among women and older workers. The session suggested that openly naming this fear can accelerate adoption.
The conference floor showcased practical AI applications. Agilysys highlighted that siloed data prevents personalization; a guest’s nut allergy entered in one system can be surfaced to front‑desk, spa, and golf shop automatically. Infor demonstrated an overnight agent that feeds live flight‑delay data to housekeeping, prioritizing rooms for late arrivals. Trybe’s overnight engine automatically splits spa‑break bookings across multiple property systems, with pilot properties reporting over £150 000 in revenue within two months.
Actabl addressed AI hallucination by training models on data relationships rather than raw spreadsheets, allowing the system to generate accurate queries. The company stopped at fault diagnosis for boilers, noting that incorrect calls could be costly.
Distribution‑side vendors focused on answering unanswered calls. Canary’s voice AI can handle 30 % of inbound hotel calls, leaving a manager a callback list if needed. Cendyn noted that no hotel can book directly inside ChatGPT yet; instead, bookings still occur on hotel websites or OTAs. The company’s new tool scans assistant responses for consistency with the hotel’s website, flagging discrepancies.
The conference underscored that clean, connected data is the bottleneck for AI, not the models themselves. It also left unresolved who bears the cost when a booking is made through an AI channel.
In summary, HITEC 2026 presented concrete AI use cases that improve operational efficiency, while also highlighting significant challenges—energy consumption, data integration, brand control, and workforce implications—that the industry must address.
In the “Distribution 2.0” session, Fredrik Sjoberg of Maison, an AI‑concierge platform, argued that by 2030 most hotel bookings will be made by autonomous agents rather than human travel agents. He noted that human agents historically captured about 10 % of booking revenue, OTAs 15–25 %, and app stores 30 %. Sjoberg cautioned that distribution costs are unlikely to fall and urged hotels to make content machine‑readable, including detailed pet policies, and to seek machine‑connectable endpoints.
The afternoon panel, “The Dark Side of AI,” focused on the environmental and cognitive costs of rapid AI adoption. Panelists cited a 2025 IEA report that AI data‑center electricity rose 17 % year‑on‑year, with most power still coming from coal and gas. Carl Winston of San Diego State, who certifies sustainable resorts, said that AI’s energy footprint complicates sustainability claims. The panel also referenced a study showing that users who let AI write content had poorer recall of the material.
Brand survival was debated in a headliner session. Lennert de Jong of Another Star and Scott Strickland of Wyndham argued that loyalty programs remain effective because AI can become a booking channel that still rewards points. Floor Bleeker countered that if agents talk to agents, hotels may lose distribution control and must compete on guest experience instead. The discussion also touched on whether hotels need staff at all; Bleeker suggested staff can be a barrier to service, while Keryn McNamara of Aimbridge insisted that hospitality is fundamentally a human business.
On day two, James Taylor emphasized that AI should augment rather than replace human curiosity. He cited research indicating that fear of appearing less capable deters use, especially among women and older workers. The session suggested that openly naming this fear can accelerate adoption.
The conference floor showcased practical AI applications. Agilysys highlighted that siloed data prevents personalization; a guest’s nut allergy entered in one system can be surfaced to front‑desk, spa, and golf shop automatically. Infor demonstrated an overnight agent that feeds live flight‑delay data to housekeeping, prioritizing rooms for late arrivals. Trybe’s overnight engine automatically splits spa‑break bookings across multiple property systems, with pilot properties reporting over £150 000 in revenue within two months.
Actabl addressed AI hallucination by training models on data relationships rather than raw spreadsheets, allowing the system to generate accurate queries. The company stopped at fault diagnosis for boilers, noting that incorrect calls could be costly.
Distribution‑side vendors focused on answering unanswered calls. Canary’s voice AI can handle 30 % of inbound hotel calls, leaving a manager a callback list if needed. Cendyn noted that no hotel can book directly inside ChatGPT yet; instead, bookings still occur on hotel websites or OTAs. The company’s new tool scans assistant responses for consistency with the hotel’s website, flagging discrepancies.
The conference underscored that clean, connected data is the bottleneck for AI, not the models themselves. It also left unresolved who bears the cost when a booking is made through an AI channel.
In summary, HITEC 2026 presented concrete AI use cases that improve operational efficiency, while also highlighting significant challenges—energy consumption, data integration, brand control, and workforce implications—that the industry must address.
