Radisson Hotel Discovery in ChatGPT Challenges the Traditional Booking Funnel

Radisson Hotel Group has brought hotel search into ChatGPT, and its first two months produced a notable result. The Radisson hotel discovery experience converted visits into bookings at roughly 1.5 times the rate of the company’s organic search traffic.

That figure comes from July and August 2026, so it is an early company-reported comparison rather than a settled industry benchmark. Still, it gives the launch more weight than a routine chatbot announcement. Radisson is testing whether a conversational answer can become a more productive entrance to its direct booking channel than a conventional search result.

The company worked with Accenture Song and OpenAI to build the experience in six weeks. Travelers can explore Radisson properties, compare rates, inspect amenities, and view locations on a map without leaving their ChatGPT conversation. Payment still happens on Radisson’s website.

The strategic contest is not simply Radisson versus another hotel chain. It is the hotel’s direct channel versus the established discovery funnel controlled by search engines, online travel agencies, and metasearch services.

Radisson wants to meet travelers before they assemble a shortlist elsewhere. Yet that ambition depends on ChatGPT surfacing the plugin reliably, travelers trusting its recommendations, and Radisson proving that early conversions represent durable demand.

Radisson Hotel Discovery Moves the Shortlist Into ChatGPT

The main change is where travelers can begin evaluating hotels, not where they complete payment.

The Radisson launch connects ChatGPT with hotel information supplied through Radisson’s systems. The plugin can show property suggestions, prices, amenities, hotel details, and map-based results inside the conversation.

A traveler might ask for a hotel near a business district, with a gym, that also suits a weekend extension. Another might combine location, family needs, and preferred amenities in one request. That structure is closer to a conversation with an adviser than a sequence of filter selections.

The plugin covers Radisson Hotel Group’s portfolio across Europe, the Middle East, Africa, and Asia Pacific. OpenAI says the group has more than 1,640 hotels operating or under development across ten brands.

That broad portfolio matters because conversational discovery works best when a system can respond to several constraints. A narrow inventory might fail whenever the traveler changes location, budget, or trip type.

The plugin does not independently invent current rates or availability. It connects with Radisson’s existing systems and presents structured results through ChatGPT. This distinction reduces the gap between a general AI suggestion and an offer that a traveler can actually consider.

Once a traveler selects a property, the booking path continues on Radisson’s website. The current experience therefore supports discovery, comparison, and referral rather than a complete in-chat transaction.

That boundary is important. The phrase AI hotel booking can suggest that an assistant handles payment, confirmation, and post-booking service. Radisson’s present implementation stops before those stages.

The company’s initial July launch required users to start prompts with @RadissonHotels. A later ChatGPT update removed that requirement for people who already installed the plugin. ChatGPT can now surface it automatically when the system considers it relevant.

This change lowers a meaningful adoption barrier. Most travelers will not remember a special invocation every time they plan a trip. Automatic selection gives the plugin a chance to appear during ordinary travel conversations.

However, automatic selection also transfers influence to ChatGPT. Radisson controls its inventory, content, rates, and booking site, but the conversational platform influences whether that information enters the traveler’s consideration set.

That dependency creates the article’s central tension. Radisson has established a direct path through ChatGPT, but it does not fully control the entrance to that path.

The launch also combines organic discovery with sponsored advertising inside ChatGPT. Radisson is treating the conversational interface as both a product surface and a marketing channel.

That combination resembles search marketing in one respect. A company can pursue unpaid visibility while also buying placement. The difference is that both routes can appear within an answer that feels like personalized advice.

For travelers, the immediate benefit is reduced coordination. They can describe a trip in normal language, revise requirements, compare properties, and then move to a live booking page.

For Radisson, the benefit is earlier access to intent. The company can reach a prospective guest while that person is still translating a general trip idea into a hotel shortlist.

Why the Radisson ChatGPT Plugin Targets Direct Bookings

Radisson is using conversational discovery to defend the direct customer relationship before another platform captures the decision.

Hotels have long balanced reach against control. Online travel agencies can expose properties to large audiences, while direct channels let hotel groups manage the booking relationship and gather first-party customer signals.

Search engines occupy another key position. They help travelers move from broad questions toward maps, reviews, comparison pages, and booking links. A hotel website often enters the process after several intermediaries have shaped the shortlist.

ChatGPT compresses those discovery steps. A traveler can ask one compound question that includes destination, dates, amenities, party composition, and trip purpose. The answer can narrow the market before the traveler visits any hotel website.

That compression creates urgency for hotel brands. If an AI assistant recommends three properties, every property excluded from the answer loses visibility earlier than it would in a traditional results page.

Accenture described the shift in similar terms in its launch details. Prospects increasingly gather inspiration, compare alternatives, and approach a decision before reaching a hotel site.

Radisson’s response is to make its own inventory usable within that environment. The company is not waiting for ChatGPT to summarize scattered webpages or third-party listings. It is supplying structured information through a dedicated integration.

The Radisson ChatGPT plugin also directs the final transaction to the group’s website. That design gives the company a chance to preserve the direct booking relationship after ChatGPT handles the earlier conversation.

OpenAI reports that the plugin’s visit-to-booking conversion rate during July and August was approximately 1.5 times Radisson’s organic search rate. This comparison suggests that users arriving through the plugin had relatively strong intent.

The result does not prove that ChatGPT is universally better than organic search. The two audiences might differ substantially. Early plugin users may be curious, highly motivated, or more willing to experiment with a guided booking path.

The denominator also matters. OpenAI disclosed the relative conversion rate but not total plugin visits, completed bookings, average booking value, or cancellation rates. A high rate from a small or unusually qualified cohort can still produce limited business volume.

Even with those limits, the result identifies a plausible mechanism. Conversational users can refine their preferences before clicking through, so the traffic reaching Radisson’s site may be more qualified.

A conventional search visit may come from a broad query, a branded query, or a traveler collecting preliminary information. The plugin can gather several constraints before showing a property and booking path.

That difference could improve conversion without making ChatGPT a larger acquisition source. Conversion efficiency and channel scale are separate questions, and Radisson has only supplied clear evidence about the first.

The direct-channel strategy also extends beyond immediate sales. A booking on Radisson’s site can support loyalty recognition, reservation management, customer service, and future direct communication.

Radisson and Accenture say they plan to add deeper personalization, loyalty recognition, in-chat booking, reservation management, trip modification, and an AI concierge. Those items are plans, not current capabilities.

If delivered, they would move the integration from referral toward account-based service. The platform could then influence more of the relationship, from initial inspiration through changes made after confirmation.

That expansion would also raise new questions about permissions and data handling. Personalized recommendations require more context than an anonymous hotel search, especially when loyalty history or prior trips enter the conversation.

For now, the business case is narrower and easier to evaluate. Radisson wants qualified discovery traffic that reaches its own booking channel before another intermediary completes the traveler’s shortlist.

ChatGPT Travel Planning Creates a New Gatekeeper

The direct booking path removes some traditional friction while giving ChatGPT more influence over which hotel brands receive attention.

Radisson built the plugin using infrastructure based on the Model Context Protocol, or MCP. MCP is an open standard that lets AI systems connect to external tools and data sources.

OpenAI’s Apps SDK architecture extends that standard with interactive interfaces and application logic. It allows developers to connect ChatGPT with live backend systems instead of relying only on model-generated text.

Accenture created an MCP server and supporting APIs for Radisson. According to OpenAI, that infrastructure powers both the plugin and parts of the company’s advertising experience.

This shared foundation is strategically useful. Radisson can structure hotel content, rates, availability, and booking signals once, then expose them through multiple conversational placements.

The six-week development period suggests that established companies can create a focused ChatGPT experience without rebuilding their core reservation systems. The interface sits above those systems and translates conversational requests into structured actions.

That speed changes the competitive timetable. Hotel groups no longer need to wait for a complete redesign of their websites or mobile apps before experimenting with conversational commerce.

Yet the technical connection does not guarantee discovery. A traveler must have access to the relevant ChatGPT feature, connect the plugin when required, and make a request that causes the platform to use it.

ChatGPT can now surface an installed Radisson plugin without the @RadissonHotels command. That improves usability, but the platform still determines relevance.

This makes ChatGPT a new gatekeeper. In traditional search, brands monitor rankings, advertising positions, map results, and referral traffic. In conversational discovery, they must also monitor whether an assistant selects their connected service.

That selection process is less visible than a ranked results page. A traveler may receive a concise answer without seeing every eligible source or understanding why one integration appeared instead of another.

The concern is not theoretical. Skift’s independent app testing found that connected travel apps were sometimes ignored or incorrectly described as unavailable by ChatGPT.

Skift tested Booking.com, Expedia, and Viator integrations. The apps eventually worked after additional prompting, but the inconsistent activation exposed a difference between technical availability and practical visibility.

That finding places Radisson’s automatic surfacing update in context. Removing a typed command helps only if ChatGPT recognizes the relevant intent and invokes the plugin dependably.

The same issue creates a measurement challenge. A hotel group must separate failures in its own inventory or interface from failures in the platform’s selection behavior.

The platform relationship also changes brand presentation. Radisson can design maps, property cards, and data connections, but ChatGPT frames the surrounding conversation.

That framing can influence the criteria travelers emphasize. An assistant might prioritize distance, amenities, review information, or another signal when narrowing options. Brands will want to understand how those priorities affect inclusion.

The pressure extends beyond hotels. Search engines and online travel agencies have spent years building comparison tools, advertising markets, loyalty systems, and extensive supply networks.

Conversational interfaces do not erase those advantages. They provide a new layer that can draw upon the same inventory while changing how travelers express intent.

Expedia, for example, already offers dynamic travel results inside ChatGPT. Its app returns flights and lodging with prices and availability, then sends users to Expedia for payment.

Booking.com was also among OpenAI’s early app partners. These companies can offer broader inventories than a single hotel group, making them useful when a traveler wants to compare many brands.

Radisson’s narrower catalog has a different advantage. It can present first-party hotel information and lead the traveler directly into its own booking environment.

The resulting contest is not a simple replacement story. ChatGPT can surface a hotel brand, an online travel agency, or another connected service. Each route competes to become the source that interprets the traveler’s request.

Early Conversion Gains Do Not Settle the Booking Question

Radisson’s conversion result is encouraging, but the available evidence cannot yet show whether the channel will scale efficiently.

The 1.5-times figure compares plugin visits with organic search during July and August 2026. It offers a concrete signal, but it leaves several analytical gaps.

First, OpenAI and Radisson have not disclosed traffic volume. A relatively small group of early adopters might behave differently from the broader population that uses hotel search.

Second, the comparison does not reveal the query mix. Organic search includes branded, nonbranded, informational, local, and navigational visits. Each category can have a different probability of producing a booking.

Third, the figure does not disclose average booking value. A channel can convert more visits while producing shorter stays, lower room rates, or different property mixes.

Fourth, the public account does not provide cancellation or modification rates. A completed checkout does not always translate into a realized stay.

Fifth, the measurement period was short. July and August include seasonal travel patterns, and the plugin was new enough to attract users who actively wanted to test it.

Those limits do not invalidate the result. They define what it can support. The evidence shows that early plugin traffic converted at a higher reported rate than Radisson’s organic search traffic during a specified period.

The advertising data needs similar care. OpenAI reports that 54 percent of recorded checkout and booking events from Radisson’s campaigns were attributed through view-through measurement.

View-through attribution credits an advertisement that a user saw before converting, even without a direct click. It can capture advertising influence that click-only measurement misses.

However, it does not establish that the ad caused every credited booking. Some users might have booked after encountering Radisson through another path.

Radisson’s use of both sponsored placements and an organic plugin complicates that picture. A traveler might see an advertisement, later receive a plugin result, and then book directly.

That sequence can be commercially valuable even when attribution remains uncertain. It also makes clean channel comparison harder.

There is a broader trust issue. Hotel selection depends on current rates, room rules, cancellation terms, taxes, location, and property details. A conversational summary must not obscure conditions that influence the final decision.

Expedia’s own ChatGPT guidance warns that AI-generated experiences can contain errors or outdated information, even when the underlying app supplies dynamic data. It advises travelers to verify details before booking.

Radisson’s direct connection should improve freshness for its own inventory. It cannot eliminate every risk created by conversational interpretation, interface changes, or incomplete user requests.

The handoff to Radisson’s website provides a useful checkpoint. Travelers can review official terms before paying, while Radisson retains its established transaction process.

That safeguard also exposes the current limit of AI hotel booking. The conversation reduces discovery work, but users still cross into a conventional checkout flow.

The planned addition of in-chat booking would raise the stakes. A conversational system handling a transaction must present price, taxes, cancellation terms, loyalty benefits, and consent clearly.

Reservation modifications add another level of complexity. A mistaken hotel suggestion is inconvenient, while a mistaken date change or cancellation can produce immediate financial consequences.

Personalization introduces a related tradeoff. Loyalty recognition and trip history could improve recommendations, but they require explicit controls over what data enters the conversation.

OpenAI says connected apps must publish privacy policies, minimize data collection, and explain permissions. Those rules establish a baseline, while implementation quality will determine whether travelers understand the exchange.

Radisson therefore has two proof obligations. It must show that the channel can deliver meaningful booking volume, and it must preserve accuracy and trust as the experience handles more consequential actions.

Hotels Now Compete for Answers, Not Just Rankings

Conversational travel shifts competition from occupying a search position to becoming the service an AI assistant chooses to use.

The older digital funnel gave hotels several visible surfaces. A brand could optimize webpages, bid on keywords, manage map listings, cultivate reviews, and negotiate distribution through online travel agencies.

Those channels remain important. The Radisson hotel discovery launch adds another contest rather than removing the existing ones.

The conversational contest starts with structured information. A hotel brand needs current rates, inventory, amenities, location data, and booking paths that an AI system can retrieve and present reliably.

It also needs a clear reason for the assistant to invoke its service. Radisson offers first-party access to its own portfolio, but Expedia and Booking.com can compare a wider market.

That creates a meaningful tradeoff for travelers. A brand-specific tool can provide direct information and a direct booking path. A marketplace can provide breadth across many brands.

Hotel groups with strong loyalty programs may emphasize personalized recognition and direct benefits. Marketplaces can emphasize selection, packaging, and cross-brand comparison.

Search engines face a different pressure. If users complete the discovery and shortlisting stages inside ChatGPT, fewer hotel decisions begin with a conventional list of links.

That does not guarantee lower search traffic across the market. Travelers may still verify neighborhoods, reviews, attractions, transportation, and safety through search.

The shift concerns who organizes the decision. A conversational system can synthesize multiple needs before displaying options, reducing the traveler’s need to assemble information manually.

For hotel marketers, that means traditional keyword rankings tell only part of the story. Teams will also need to test prompts, inspect plugin activation, monitor data quality, and compare conversational referrals with other traffic.

The required expertise crosses organizational boundaries. Ecommerce teams understand conversion and direct booking. Data teams manage inventory signals. Marketing teams handle acquisition and attribution. Legal teams review permissions and claims.

A conversational product connects all four. A stale amenity field can affect a recommendation. An unclear booking path can damage conversion. A weak permission screen can undermine trust.

The Radisson ChatGPT plugin is particularly notable because its infrastructure also supports advertising. That connects product development with media buying.

The pairing might help Radisson learn which prompts generate demand and which property details influence selection. It could also make organic and paid performance difficult to separate.

Competitors now face a practical decision. They can build their own integrations, rely on marketplaces, optimize public content for AI answers, or combine those approaches.

A dedicated integration offers more control over live data and user experience. It also requires maintenance across changing platform rules, interface patterns, and technical standards.

Relying on a marketplace reduces that burden but keeps an intermediary in the booking path. Optimizing public content may improve visibility without ensuring access to live rates or inventory.

Radisson chose the dedicated route while retaining its existing direct site. Its early result suggests that this route can attract high-intent visitors, but the company has not shown that it can become a major source of total bookings.

The channel’s economics will matter as competition grows. If every large hotel group builds a plugin, ChatGPT must decide which services to recommend for a general hotel request.

Sponsored placements add another variable. A user may encounter an organic app result, a paid hotel placement, or a marketplace result within the same planning process.

That mix resembles the commercial tensions of search, but the interface is more conversational and selective. The answer can feel like advice even when paid distribution contributes to visibility.

Platforms will need clear labeling and consistent app-selection behavior. Brands will need reporting that distinguishes organic invocation, paid exposure, referral visits, and completed bookings.

Without that clarity, companies may struggle to tell whether they built a productive new channel or simply moved marketing spend into a less transparent interface.

What Will Prove Radisson’s AI Hotel Booking Strategy

Three signals will determine whether Radisson has opened a durable booking channel or produced a promising early experiment.

The first signal is sustained performance at larger scale. Radisson should disclose enough data to show whether the conversion advantage survives beyond the July and August launch period.

Useful evidence would include referral volume, completed bookings, cancellation-adjusted stays, and performance across markets. A relative conversion figure becomes more informative when readers can understand the size and composition of the channel.

If the 1.5-times advantage remains while traffic grows, the case for conversational discovery strengthens. If it falls sharply as the audience broadens, early adopter selection likely explained part of the result.

The second signal is reliable automatic invocation. ChatGPT’s ability to surface an installed app without a typed command is central to mainstream adoption.

Travelers should not need to know the name of a plugin before using it. They also should not need repeated prompting when an appropriate service is already connected.

Independent testing should examine whether Radisson appears consistently for relevant destinations and constraints. It should also measure whether results remain correct across devices and changing trip requirements.

Reliable invocation would show that the experience functions as a genuine discovery channel. Inconsistent invocation would keep it closer to a feature for informed users.

The third signal is the delivery of account-aware and transactional features. Radisson and Accenture have identified loyalty recognition, deeper personalization, in-chat booking, reservation management, trip changes, and concierge functions as future priorities.

Loyalty recognition would test whether the integration can move beyond anonymous search. It could let the system account for membership status, preferences, and eligible benefits.

In-chat booking would be an even larger step. It would require the conversation to communicate final terms and obtain clear authorization before confirming a reservation.

Reservation management would test operational reliability after the sale. Changing dates, canceling a room, or resolving a disruption carries more risk than displaying a shortlist.

These additions would strengthen Radisson’s claim that conversational AI can become a full direct channel. Delays or limited adoption would suggest that the handoff to established booking systems remains essential.

Travelers should watch the experience with the same caution they apply to other booking tools. Use conversation to clarify needs and build a shortlist, then verify dates, room details, taxes, and cancellation terms before confirming.

Hotel and travel companies should watch something broader. Radisson is testing whether a brand can place its own live inventory inside an AI conversation and still preserve the direct customer relationship.

That is why the Radisson hotel discovery launch matters. It is not merely a new interface for browsing properties. It is a bid to control the path from travel intent to direct booking before an intermediary defines the shortlist.

The early conversion data gives that bid credibility, not final proof. The decisive evidence will come from scale, reliable discovery, and the safe delivery of more transactional features.

Will travelers treat ChatGPT as the place where hotel decisions begin, or as another research layer before returning to familiar booking tools? Radisson now has a working channel through which to find out.