Agents, Not Interfaces: Is This the Beginning of the End for UX?
9 experts shared their view
This morning, a hotel was booked under my name. But to be clear: I didn't book it. The reservation was completed autonomously by the newly launched ChatGPT Agent, an intelligent assistant capable of interpreting natural language, learning from previous interactions, and parsing massive real-time datasets. The process was (overall) smooth, even elegant. The agent handled ambiguity, narrowed down options, and delivered increasingly relevant suggestions until the final step, when it politely asked me to input my payment details. Not surprisingly, the agent didn't explore quirky boutique websites or reward clever brand storytelling. It went directly to Booking.com, not because it was cheaper or better, but because it was the most machine-readable.
So here's the real question: if autonomous agents become the dominant gateway between users and suppliers, what happens to UX as we know it? Do we still need beautifully designed websites, optimized booking flows, and handcrafted content? Or are we moving toward a future where OTAs serve not human travelers, but intelligent agents, acting as data reservoirs, not discovery platforms?
It was inevitable that we would move to conversational AI and on-demand UI. The writing was on the wall years ago. Hotels will be caught off guard as OTA AI agents beat them without mercy at the pricing game. It’s time to build a defensive tool.
- We will still need beautifully designed websites and flows — but more as validation layers and brand showcases than as discovery engines.
- OTAs and suppliers will increasingly serve as structured data ecosystems rather than pure UX-driven platforms.
- The real battleground won"t be "prettier booking paths," but better data standards, APIs, semantic precision, and trust architecture for agent-mediated decision-making.
Are we moving to a future - interesting... Though, you remove glossy images and brand (the websites of today), don't we just return to the flat green text screens of travel agents past, but this time operated by AI? Where description and detail was the important asset to capture interest and convert the booking?
Where there is a world in which people want everything automated and instant, there is also a world where people want more tangible experiences. For many, it's the journey to the booking that counts more than the booking itself. Isn't that, after all, the mantra of travel - the journey, not the destination?
Without evoking people's senses, where and how do you build anticipation and excitement? And must we all remember the days when we blindly arrived at a destination to be ferried to our hotel in the middle of the night - only to wake to the horror of reality?
Tech needs to talk to tech, but people need to talk to people for true experiences.
I am of the view that this is a very real outcome in the future, perhaps only the rudimentary beginning.
In the grand scheme of things, if you think about the GDS and the standard interfaces designed to normalize data and make everything somewhat homogenous, then it doesn’t take much imagination to remove the agent and replace it with a technology agent equally unconcerned with pretty pictures and sensory stimulation.
That said, AI has learnt/is also learning to understand imagery. Who is to say it will not be able to parse value from both rich imagery and homogenous data, making decisions that are less analytically pure but more attuned to an end user’s personal likes and feelings?
Even better, after booking, it could provide a storybook of the potential journey it has mapped out based on the traveler’s instructions.
None of this is too much of a stretch for the imagination! The real question will be how much a person values the discovery process of travel shopping.
Autonomous AI agents are changing how travelers interact with digital platforms. At Amadeus, we see this shift not as the end of UX, but the start of a new kind — one designed for both humans and machines.
As agents take over tasks like booking, the focus moves from visual design to data clarity. Websites and platforms must be easy for machines to understand, not just people. That means using structured data, smart APIs, and fast access to relevant content.
But storytelling and design still matter. People trust brands that feel human, and agents learn from those human preferences. The future of UX will blend creativity with machine-readability — helping agents make better choices while keeping travelers inspired.
In short, UX isn't disappearing. It's evolving to serve a new kind of user: intelligent agents acting on behalf of real people.
Yes, we are moving toward autonomous booking, whether through OTAs or other channels. Eventually, booking flows will be managed by agents, which means the traditional focus on optimising the flow itself may diminish.
However, does it make website UX and content completely irrelevant? No, on the contrary, they remain important.
Different AI platforms process and prioritise information in different ways, some relying heavily on structured, real-time data, while others evaluate trust scores, reputation, sustainability, or compliance. Many draw primarily from OTA content.
It means continuing to ensure your content is AI-ready is essential. That does not just mean being search-optimised but also generative engine optimised (GEO). The data needs to be structured in a way that AI can interpret it logically, respond accurately to queries, and match your property as the right fit.
Consistency is equally important: website and OTA content should remain fresh, aligned, and updated. Any discrepancies risk sending confusing signals to the AI, which could reduce visibility or lead to missed opportunities.
In short, UX as we know it may change, but content quality, consistency, and machine readability will matter more than ever. What we need to do is be ready to adapt and evolve with AI.
Thank you, Simone, for sharing a powerful example of how intelligent agents are reshaping hotel booking. It raises an important question: for whom should hoteliers design their websites?
In this evolving landscape, we must now design for two audiences: humans and agents. For travelers, traditional UX elements, such as visuals, navigation, and storytelling, remain essential. These essential elements will not disappear anytime soon. For agents, what matters the most is structure and machine readability. A website must be accessible to algorithms that parse, compare, and recommend properties.
In practice, this change means websites must now support both human–machine and machine–machine interactions. Assuming most hospitality marketers have already mastered UX for people, the next challenge is to also support Agent Experience (AX) --- ensuring data clarity and accessibility for intelligent systems while maintaining inspiration for humans.
So rather than the "end" of UX, we may be entering an era where the real design competition is not for human eyes, but for machine interpreters. The winners will be the hotels or OTAs whose stories resonate not only with people, but with the algorithms deciding what people see first.
Autonomous agents are transforming the way travel bookings occur, marking a significant evolution in the hospitality industry’s digital user experience (UX). Traditional UX has have been designed to captivate human guests, emphasizing intuitive navigation, immersive storytelling, and emotional engagement. Now, intelligent agents are emerging as primary intermediaries, with priorities shifting toward data accessibility, standardization, and machine-readability.
This transition does not diminish the value of UX; rather, it broadens the scope. To stay competitive, hospitality providers must ensure their property details, rates, and inventory are structured for effortless integration and accurate interpretation by automated platforms. Adopting emerging standards such as the Model Context Protocol (MCP) can facilitate this process, allowing seamless communication with AI-driven systems through clear APIs, comprehensive metadata, and consistent organization.
That said, the human element in hospitality remains indispensable. Guests will continue to seek personalized recommendations and authentic experiences, especially for more complex or memorable journeys. Therefore, the sector should embrace a dual-track approach: developing “agent-ready” systems for efficient, data-driven interactions—leveraging MCP and similar protocols—while maintaining engaging, human-centered designs for those moments where personal connection and brand identity make the greatest impact.
Ultimately, the future belongs to organizations that can effectively blend digital efficiency with the enduring strengths of human hospitality.
AI Search bots pull only 25% of AI answers from hotel website content (VertoDigital). All of the generic hotel information, descriptions of services and amenities the AI Agent platforms have already scraped from publicly available information about the hotel on the Internet or from proprietary databases.
Agentic AI does not need websites for information or to make a booking. To access hotel ARI (Availability, Rates and Inventory), Personal AI Agents will communicate directly via APIs with the hotel cloud PMS, CRS, Channel Manager, the future hotel's own AI Agent, or via MCR middleware or APIs with the OTAs.
Personal AI Agents like ChatGPT Operator, Google Gemini, Microsoft Copilot, Claude AI Agent, etc. are already a fact.
In the near future, the Personal AI Agents will have the option to research, plan and book travel by making a "handshake" with the newly emerged hotel own AI Agent, or with the OTA's own AI Agent. If hoteliers do not provide the option, the only choice will be the OTA Agent.
I believe that in the future Agentic AI and AI Agent platforms will make hotel websites and mobile apps completely obsolete since AI does not need hotel websites for information or ARI.