HITEC Paris Panel Warns Hotels Face a Second Intermediation Crisis

Only about one in ten hospitality businesses in the Netherlands currently uses artificial intelligence, according to research presented by Dr. Alexander Schmidt of Hotelschool The Hague at the inaugural HITEC Paris. The figure framed a day-long debate about whether hotels are using technology to improve hospitality or quietly redefining hospitality around what technology can do.

HITEC Paris ran for the first time as part of EquipHotel, organized by RX Global in partnership with Hospitality Financial and Technology Professionals (HFTP). The opening AI panel was moderated by Lyle Worthington of HFTP and featured Schmidt, Fabio Maggioni of Belmond, and Simone Puorto of Hospitality Net, Travel Singularity and Elegia.

Who controls the intelligence layer?

Schmidt warned that AI could insert a new layer of intermediation between hotels and their guests, comparable to the shift OTAs (online travel agencies) forced on distribution decades ago. Big technology companies are pouring capital into foundation models and will need to monetize those investments. Hotels risk paying new fees and, more dangerously, losing ownership of the guest relationship yet again.

The question is no longer only who controls distribution but who controls the intelligence layer through which travelers discover, evaluate and book accommodation. If AI assistants become the primary consumer interface, hotels may negotiate with intermediaries that control not just visibility and transactions but the recommendations shaping decisions.

From buying software to owning the data

Maggioni drew on his work building proprietary AI agents at Belmond to challenge a longstanding industry assumption: that hotel companies should operate hotels and leave software to vendors. Ownership of the data layer matters more.

"The key is to own your data layer. When your information lives in your own platform, you can plug in any third-party system, move faster, and deploy your own AI agents without waiting for vendors," Maggioni said. The goal is not large engineering departments; small multidisciplinary teams can build and test solutions rapidly.

One prerequisite is a coherent stack. Hotels rushing to deploy AI agents on top of fragmented property management, reservation, CRM and operational systems that barely communicate may simply make inefficient systems more expensive. The panel questioned whether every hotel needs an AI strategy at all. Open interfaces, flexible integrations and clear data ownership may prove more durable than any single AI application.

Three dimensions, one verdict

Schmidt proposed evaluating AI investments across three dimensions: operational efficiency, employee experience and guest experience. Vendors emphasize the first. The other two change with every deployment.

"Every AI decision affects three dimensions: operational efficiency, worker experience, and guest experience. The technology matters, but the strategic position you choose matters more," Schmidt said.

The distinction matters most for what operators do with time saved by automation. A hotel can redirect freed hours toward guest relationships or toward headcount reduction. Both can be economically rational. They are not the same understanding of hospitality, and suitability depends on brand positioning.

Humans measured like machines

The panel flagged a paradox: the industry celebrates AI for becoming more human while measuring employees on increasingly mechanical KPIs. A thirty-second check-in is deemed efficient; a conversation is not.

"We're training humans to work like machines and evaluating them on technical KPIs. Until we start measuring humans on human KPIs and machines on machine KPIs, rather than the other way around, we'll keep missing what makes hospitality truly human," Puorto said.

Schmidt cited AI-enabled employee monitoring that tracks how often staff use prescribed expressions. Such systems can improve measurable output while eroding autonomy and authenticity.

Luxury, scarcity and the squeezed middle

For luxury, Puorto's Humans-as-Luxury framework argues that human experiences gain scarcity value precisely because automation makes them optional. An imperfect, spontaneous interaction may differentiate a property in an algorithmically predictable market. Maggioni countered that luxury guests expect recognition, continuity and effortlessness — which intelligent systems can deliver when they support staff rather than replace them.

The harder problem is the midscale segment. Midscale hotels face the greatest technological fragmentation, limited investment capacity and margin pressure. They risk a difficult middle ground: unable to capture the full benefits of automation, and unable to afford the staffing that makes human service a differentiator.

When the guest is an agent

Schmidt suggested that AI-mediated discovery could rewrite hotel marketing economics. Autonomous agents evaluating properties on a traveler's behalf will prioritize structured information, transparent pricing, availability and reviews over visual storytelling. Puorto's prescription: build a single, reliable source of truth covering descriptions, amenities, menus, policies and guest preferences. In agent-to-agent commerce, data quality becomes a prerequisite for visibility.

The panel also weighed digital sovereignty. Most AI hospitality products depend on a small number of foundation-model providers, with uncertain future pricing and access terms. And companies sometimes deploy agents where a conventional integration would be cheaper. Sometimes an API is still a better answer than an AI agent.

The employment question

Worthington closed on automation's central contradiction: "Are we freeing people up, or are we just trying to remove them? That's ultimately the question we need to ask ourselves when we talk about automating hospitality."

Panelists flagged the potential loss of entry-level roles, which traditionally gave young professionals the judgment and practical experience needed to run complex organizations. Adaptability, Worthington argued, may matter more than mastering any single application.

The panel's clearest principle: the ability to adopt technology is not the same as having a reason to adopt it. Whether more artificial intelligence also means better hospitality remains a choice — and one the industry must make before 2030.