What It Takes to Make AI Booking Work at Scale

Skift Take The travel industry is largely still testing AI bookings in controlled environments. Companies that can prove the technology works across live bookings at scale could gain a valuable head start. This sponsored content was created in collaboration with a Skift partner. Most AI booking tools entering the market have yet to prove themselves at scale. Deloitte’s 2026 State of AI in the Enterprise report found that only 25% of organizations had moved 40% of their AI pilots into production, while 37% were still using AI at a surface level without materially changing underlying processes. HotelPlanner came to market from a different starting point. The company developed Reservations.ai, a voice-driven, fully integrated booking system, to handle its own multi-supplier marketplace before ever considering it as a standalone product. Only now is it beginning to open the technology to outside partners, after testing it against live bookings across flights, hotels, and other travel inventory. Skift Studio spoke with Tim Hentschel, co-CEO of HotelPlanner, to explore what it takes to make AI booking technology work at scale, and what travel companies should consider when deciding whether to build or buy. Tim Hentschel, co-CEO, HotelPlanner Skift Studio: What problem was Reservations.ai built to solve, and when did you see its potential as a product for other travel companies? Tim Hentschel: Reservations.ai started during the travel industry’s post-pandemic recovery. At the time, group travel was a core part of HotelPlanner’s business, but COVID-19 had brought that market to a halt. Group sales managers and hotel sales directors were suddenly sitting at home with very little to do, so we saw an opportunity to put their extensive local hotel knowledge to work. We expanded our traditional outsourced call center into a gig-powered model, bringing in people who really knew the hotels. We recorded 10 million highly rated calls and began using them to train an AI-powered, fully voice-activated reservationist around 2023. About a year and a half ago, we got it to the point where it could complete hotel reservations at the same level as a human. The next logical step was a fully optimized AI travel agent capable of handling air, car, hotel, and event tickets. That became Skyler, the Reservations.ai’s AI-powered travel agent. Making it available to other companies was a natural progression. We’ve always had a strong B2B affiliate platform, and once the product became operational, our affiliate managers wanted to help partners solve the same problem with the technology. What can Reservations.ai do today across the booking journey, and how does it connect with live inventory, supplier systems, payments, and post-booking servicing? Reservations.ai has access to everything HotelPlanner does across its multi-supplier platform, including direct hotel inventory, millions of concert and sporting event tickets, worldwide flights, and rental cars. Anything a human agent can access, the AI can access too, but it can process those options much faster and bring in a much deeper knowledge of the local area. The goal is to give the customer the same breadth of choice they’d get from a human agent, while making the search and booking process much quicker. We can also train the AI on a client’s existing call recordings, preserving decades of institutional knowledge and the brand identity their teams have built around customer interactions. What measurable results have you seen from using the technology internally? The results were stunningly bad at first. Conversion was only around 1–2%, but we got it up to a human-level conversion rate of roughly 14–16% over about a year. A lot of that came down to improving the back-end technology so the AI could better understand how real customers speak, including accents, slang, and local vernacular. We’ve also been able to bring hold times down to virtually nothing, which makes the technology especially useful as a complement to human agents. Around 30% of customers still ask to speak to a person, so we don’t see AI replacing humans. The better approach is to use AI for overflow and peak periods rather than overstaffing a call center. That can improve both the top and bottom lines through higher conversion at lower labor costs. Where does Reservations.ai sit within the company’s wider brand portfolio, and how do you see that evolving? Reservations.ai is now at the top of our brand hierarchy. HotelPlanner remains a strong brand for hotels and group bookings, but we increasingly see the business as being about all things reservations, thanks to AI. That includes hotels, flights, cars, event tickets, and the technology that connects them. So, this is much bigger than launching a new product. We’re even in the process of changing the signage at our corporate headquarters from HotelPlanner to Reservations.ai, and you can already see that shift reflected on reservationsgroup.com. What do you see as Reservations.ai’s biggest point of differentiation in the market? Reservations.ai can take a customer through the entire reservation over the phone using AI, including taking the credit card, processing the payment, and issuing the confirmation. We designed the experience to feel as human as possible, to the point that we now disclose at the beginning of every call that customers may be assisted by both humans and AI. We’re handling around 500,000 calls a day, with roughly a 15% call-to-reservation conversion rate. That ability to complete the transaction end-to-end, at that volume, is what we believe sets the technology apart. How is Reservations.ai currently offered to external partners? We’re flexible in how we structure it. Companies can license the technology directly or work with us on a revenue-share model, depending on what makes the most sense for their business. For a distributor, OTA, or TMC deciding whether to build agentic booking technology in-house or license an existing system, what factors should inform that decision? If you’re trying to build this from scratch and you’ve never managed a call center, you’re making a massive mistake. Even if you have a fully proprietary call center and access to your own call and customer data, it could still take three or four years to build something decent. So the question is whether you want to spend years developing something that might work by 2030, or get something live in a matter of weeks that we started building in 2022 and have already tested with hundreds of thousands of real customers every day. What’s next for Reservations.ai as a product? The next step is to make it both voice- and touch-interactive. We want customers to be able to speak naturally to Skyler, but also use touch when that’s faster or more convenient. That would be especially useful in environments like a car, where you may want to make a reservation hands-free but still be able to tap the screen to steer the search. You can speak much faster than you can type, so we see voice as the primary interaction, with touch there when you need it. We’ll keep testing that with real customers and building around how they want to use it. What role do you think AI will ultimately play across the travel industry? I’ve been through almost every major technology and investment bubble over the past couple of decades, including dot-com, mortgages, blockchain, and Special Purpose Acquisition Companies (SPACs), so you start to recognize when expectations get ahead of reality. The AI hysteria probably peaked earlier this year, and people are now coming back down to earth about what the technology can realistically do. That doesn’t make AI any less important. We see it as the future, but more as a human-AI hybrid rather than something that replaces people altogether. AI is incredibly effective when it’s helping someone do their job better, but it still needs human direction. That combination is where the real value will be in travel. Contact HotelPlanner to see if Reservations.ai fits your 2027 planning. This content was created collaboratively by HotelPlanner and Skift Studio.