AI Agents Could Change Hotel Pricing: New Risk Pricing Based on Guest?

The fundamental logic of dynamic pricing in hospitality has long been familiar: demand rises, occupancy increases, room prices change. Under the same conditions of room type, date, and booking terms, customers generally see the same available price.

The entry of AI-powered booking agents into travel distribution could complicate this model.

Hospitality Net's HN Brief from August 14 raises a compelling argument put forth by hospitality.today and reconline AG: AI-powered booking agents could leverage customer urgency, search history, loyalty level, and behavioral data to add a personalized margin on top of the price set by the hotel, with this difference going to the distribution channel rather than the hotel.

The claim is significant. However, it is crucial to accurately define the level of evidence.

The use of personal data to create personalized pricing is no longer just a theoretical scenario. Research by the U.S. Federal Trade Commission (FTC) confirms this. In contrast, there is still limited independent and comprehensive evidence that AI booking agents are systematically using this model in hotel rooms.

The main issue for hotel operators begins here.

Will the room be priced, or the customer?

In classic revenue management systems, the primary variables affecting price are room type, date, occupancy, demand forecasting, time left to booking, events, and competitive conditions.

Personalized pricing, however, uses a different logic.

The system attempts to predict how much the customer is willing to pay alongside the market value of the product.

The FTC's "surveillance pricing" study published in January 2025 revealed that third-party pricing agents can use consumer data at a sensitive level.

According to the FTC, data such as location, demographics, browsing history, shopping behavior, and even mouse movements on a webpage can be used to determine the prices or promotions consumers see. The agents examined by the agency served at least 250 customers.

This finding is not directly related to the hotel industry. However, it makes the question realistic in terms of hotel distribution.

If a booking agent knows how flexible a customer's travel dates are, how many times they have searched for the same hotel, or how close they are to making a purchase, will this information only be used to "find the best hotel"?

Or could it also be used to predict the maximum price the customer is willing to pay?

What happens if a room priced at 200 euros is sold for 220 euros?

The most striking aspect of the scenario reported by Hospitality Net is who retains the price difference.

Hypothetically, let’s say the hotel has opened a room for distribution at 200 euros.

If an AI-powered agent determines that the customer's purchase intent is very high and sells the same room for 220 euros, two different business models emerge.

If the hotel receives 220 euros, it can be considered a revenue management issue.

However, if the hotel continues to receive 200 euros and 20 euros remain with the distribution agent, the discussion shifts from pricing to distribution economics.

The argument reported by Hospitality Net draws attention precisely to the second scenario. However, the extent to which this practice occurs in hotel bookings has not been verified by independent sources.

This distinction is important.

Because the industry can examine whether its contracts and technology integrations allow for such a model instead of accepting an unproven practice as a widespread reality.

Why is surveillance pricing on the agenda?

Personalized pricing is not just a hospitality discussion.

The FTC launched an investigation in 2024 into Mastercard, Accenture, PROS, Bloomreach, Revionics, McKinsey, and other companies. The aim of the research was to understand how different prices for the same product or service can be created using consumer behaviors and personal data.

The FTC's initial findings show that the systems can use both directly obtained consumer data and inferred and third-party data.

More importantly, the technology used does not have to change the price alone.

The system can present higher-priced products more visibly to certain customers or offer different promotions based on the individual.

In hospitality, this could mean more than just changing room prices.

Room ranking, package displays, upgrade offers, breakfast-inclusive pricing, cancellation conditions, or additional service offerings can also be personalized.

The airline industry's discussion serves as a warning for hotels

One of the most visible discussions on this topic in the travel sector has occurred in airlines.

In 2025, U.S. senators questioned Delta Air Lines about its plan to expand AI-powered pricing using Fetcherr technology. The senators expressed concerns that the system could be used for personalized "surveillance pricing" purposes.

However, the other side of the discussion is also important.

In its official response to the senators, Delta clearly stated that it does not use personal data to apply individual prices to customers, has not tested such a system, and does not plan to do so. The company argued that its AI-powered pricing technology is a decision support tool that uses aggregated data and provides recommendations to analysts.

This example illustrates why careful language is needed regarding AI pricing.

Dynamic pricing with AI is not the same as setting personalized prices based on personal data.

Could an AI agent be the new OTA?

The bigger strategic question for hotels is where bookings will take place in the future.

Today, guests can search on Google, go to an OTA, or use the hotel's own booking engine.

In the Agentic AI model, however, a customer could tell an assistant, "Find me a hotel with a Bosphorus view, a gym, and within my nightly budget for next week in Istanbul."

The AI system finds hotels, compares them, and can complete the booking if it has sufficient authority.

In this case, the importance of the classic website interface diminishes, while the hotel's machine-readable data becomes more important.

Hospitality Net's AI terminology guide also highlights this transformation. The transfer of hotels' live pricing and availability information to AI systems via APIs and similar protocols is seen as one of the key components of future distribution infrastructure.

The commercial question here is simple:

If AI makes the booking decision, who will control the customer relationship?

The hotel's price may no longer be sufficient

The revenue manager can today see the prices across different channels through the channel manager.

However, if AI-based new agents can add their fees or dynamic margins on top of the price, it may no longer be enough for the hotel to simply track the question, "What price did I send?"

The question, "How much did the guest actually pay?" also gains importance.

If there is an unseen difference between the two figures, three issues arise for the hotel.

The first is price transparency.

The second is brand perception. The guest may think that the price they see was set by the hotel.

The third is revenue loss. If the customer is willing to pay more but this extra payment does not reach the hotel, the value remains with the distribution agent.

Therefore, the future rate parity discussion may not be limited to just the question, "What is my price on the OTA?"

Why is direct booking gaining importance again?

The argument reported by Hospitality Net positions direct booking as the primary defense mechanism against this risk. The source itself notes that this claim should be questioned.

Still, the structural advantage of direct booking for hotels is clear.

When the transaction occurs in the hotel's controlled booking engine, there is more control over price, payment process, and guest relationship.

In a third-party channel, the hotel leaves part of this control to the distributor in exchange for visibility and demand.

AI agents may not eliminate this equation; they could simply create a new intermediary layer.

Therefore, the goal for hotels may not be to escape AI distribution, but to create infrastructure for direct bookings through AI.

What should hotels control today?

Since this transformation is still in its early stages, some basic checks are more valuable before making significant technology investments.

The price and inventory information in the hotel's CRS, PMS, channel manager, and booking engine must be consistent.

Distribution contracts should clarify how third parties can add fees or margins on top of the price set by the hotel.

Who collects guest data, who has access to it, and for what purposes it can be used is becoming a concern not only for the IT department but also for commercial management.

The room features, cancellation conditions, F&B options, location information, and service details that AI systems will read must also be up-to-date and structured.

Because being unable to be visible in AI distribution is as much of a commercial risk as being visible with incorrect information.

What does this mean for hotels in Turkey?

Distribution in Turkey, especially in internationally sought-after destinations like Istanbul, Antalya, Bodrum, and Cappadocia, is already very layered.

OTAs, tour operators, bed banks, DMCs, GDSs, and direct channels can access the same room inventory in different ways.

The entry of AI agents as a new intermediary in this structure could further complicate the chain through which the price passes before reaching the end customer.

The first step for hotels in Turkey should not be, "AI booking is coming; we must immediately purchase new technology."

First, the visibility of the existing distribution chain should be increased.

What net price does the hotel give to which channel? Where is this price being resold? What does the guest actually pay during the booking? What customer data does the reservation reach the hotel with?

The answers to these questions are not fully visible in many businesses even today.

AI may accelerate existing distribution problems rather than create them.

Stafflink's perspective

The critical point of the AI discussion in hospitality is not whether the technology recommends a hotel to the guest, but who retains the economic control of the transaction.

If the AI agent can read the guest's budget, preferences, and purchase intent better than the hotel, this information creates significant commercial value.

The question is, who will capture this value?

In our view, hotels will need to monitor not only ADR, RevPAR, and channel commissions in the coming period but also the difference between the price sent by the hotel and the total amount the guest actually paid more closely.

However, at the current level of evidence, it is premature to conclude that AI agents systematically increase prices based on personal data for hotel customers.

The risk is realistic; the prevalence has not yet been proven.

Therefore, the correct approach for hotel businesses is not to shy away from AI distribution out of fear, but to make visible at what point the price, data, and customer relationship are controlled by whom.