New concepts require definition. Budding generative AI certainly does because it’s not the same as the now ripened AI. Unlike traditional AI, generative AI basically creates something new, whereas its predecessor was confined to certain rules. Take Apple’s omnipresent voice assistant Siri: a user gives it an input and a response is spit back out. It’s the same with Amazon’s Alexa or even the search tool on Netflix: they are given prompts and kick out a recommendation.
Think of generative AI as the next evolution of AI, where the confines of what we knew are breaking down. In other words, generative AI basically takes an idea and runs with it. It creates something brand new from an initial prompt. Yes, ChatGPT (the GPT stands for generative pretrained transformer) is a prime example of that and now part of the zeitgeist. I asked it to write an ode about my dog, Olive, a Jack Russell, and in no time flat it spat out seven stanzas of rhyme that would make Longfellow blush.
Meanwhile, companies like PolyAI are making voice assist in customer service life-like. Just listen to these calls, which are so human the real human on the other end has no clue they are speaking to an AI.
It’s no wonder that NBC said it will use AI software to recreate famed sportscaster Al Michaels’ voice to deliver daily recaps of the Paris Olympic Games. A Vanity Fair article shed more light, writing that NBC trained AI to match Michaels’ delivery using his past appearances on the network. As Vanity Fair further described it: “The feature, called Your Daily Olympic Recap, will pull from thousands of hours of live coverage from the Games in Paris using a large language model. The model analyzes subtitles and metadata to summarize clips from NBC’s Olympics coverage, and then adapts those summaries to fit Michaels’ style. The resulting text is then fed to a voice AI model—based on Michaels’s previous NBC appearances—that was trained to learn the unique pronunciations and intonations of certain words and phrases.”
Maybe the Paris Olympics have rubbed off on Paris-based Accor, or vice-versa. In April, the hotel behemoth announced it had paired with Deloitte to modernize its customer services for its luxury and lifestyle brands. Through a global cloud telephony platform and advanced service-based on AI, Accor can now provide personalized customer service to its guests in more than 110 countries. The cloud AI enabled the building of a bot that gives automated answers to guest requests. Still, complex requests are handed over to a human, but the result is a mix of advanced bot and a human touch to personalize conversations. The solution was deployed to over seven contact centers, 2,000 telephony lines and Accor said it has enhanced the productivity of 600 reservation agents. The ultimate aim is to have 40% of calls dealt with via AI.
A recent Forbes article framed generative AI this way: “It’s like an imaginative friend who can come up with original, creative content.”
Conference Talk
Each year, Hospitality Financial and Technology Professionals (HFTP) puts on HITEC, a mega-event focused exclusively on hospitality technology. This year, it was held in Charlotte, N.C., where the convention center there was filled to the brim with exhibitors promoting and pushing their wares. Consider property management systems or PMSs, which are really at the heart of running a hotel efficiently. They, for instance, facilitate a hotel’s reservation management and ancillary administrative tasks. And, like third-party management companies in the U.S., there are many, all different yet all the same in providing what a hotel needs to go.
Like the PMS, or the countless other technology now servicing the hospitality industry, more and more have a layer of AI added on. As Markus Feller, CEO of Like Magic, a digital solution to support hospitality employees and guests, recently told PhocusWire, “Change is going to come thick and fast for property management systems. AI advancements and their subsequent integrations are going to blow a lot of minds—especially when it comes to automation.”
Meanwhile, the test case for AI in revenue management seems boundless, with AI being able to make pricing decisions based on myriad inputs and do it in a conversational format. Imagine a human revenue manager asking AI what to price a hotel suite, on November 6, in Baltimore, the day before Game 7 of the World Series. AI can take data from many different sources—past visits, events and more—and deliver recommendations.
During HITEC, one panel of AI experts looked to answer some questions and clear up some misconceptions about AI in hospitality. First, as Matt Schwartz, CTO of management company Sage Hospitality Group, pointed out: “AI is everywhere and you probably use it more than you think.” That’s true: the technology has been around since as far back as the 1950s, but the focus remains on how AI can contribute to how humans can live and work better. “We focus on how technology enables the guest and employee experience,” said Shane O’Flaherty, the global director of travel, transportation & hospitality for Microsoft, whose Copilot is a generative AI chatbot that allows users to search for specific information, generate text, such as emails and summaries, and create images based on text prompts.
AI, O’Flaherty posited, would also profoundly impact how customers search and ultimately make buying decisions. “Websites and OTAs are using AI at the top end of the funnel,” he said, “and now down. People want the web experience to be more human.”
The top of the marketing funnel represents shoppers in the awareness stage, while the bottom of the funnel represents shoppers in the purchasing stage. “It used to be search and inspiration only,” O’Flaherty said.
Regarding the use of AI to field calls and customer requests, Bryan Steele, managing director of digital consultancy Jireh-Tek Limited, said that the lofty percentages companies aim for in using AI to service customers is bold but achievable. Furthermore, he said, AI also increases efficiency where it most counts: profitability. When using AI in call centers, he said he has seen as much as $180 in savings pertaining to cost per agent per day. “Revenue is up, there is lower turnover—it’s a cheaper cost,” he said.
The question, Schwartz said, boils down to this: Is AI making things better, easier? One instance, he shared, supported the effectiveness of AI. “Our sales and marketing team came to us and said that group sales needed better response times,” he said. What Sage did was put RFP responses into a single folder on Box, a cloud-based content management, collaboration, and file sharing tool used by many companies, which has its own AI solution. Now, Schwartz said, when RSVPs come in for an event, the AI functionality looks at them for the specific hotel and automatically prepares responses for sales teams to review, saving hours of work.
Another example he cited concerned a conference at a 600-room property, where, through Sage’s own ChatGPT, called Sage Chat, they were able to send out customized welcome letters by feeding it information about attendees.
“It’s the eager intern,” Schwartz said of AI. “It’s awake 24/7 and fully caffeinated and, just like an intern, you will trust, but always verify,” like when an intern creates a PowerPoint deck, you still check the work before presenting it.
He also pointed to a new AI application that allows hoteliers to inspect a room to make sure it’s properly equipped and satisfied from a standards perspective, such as ensuring the right amount of pillows are on the bed, that the room has the requisite number of water bottles and, even, that a coffee pot or kettle is facing the correct direction. “You just hold up your phone and it will tell you if everything is in order the way it should be,” he said. It’s not AI supplanting housekeeping, it only ameliorates the process.
Aside from the functional, AI will allow better and more access to data through its business intelligence application. “How do we allow everyone [internal] access to all our financials, all the time,” Schwartz said, things like knowing a property’s NOI in real time and if it was hitting on the top line and bottom line. “The idea is to unlock the data; that’s the power of AI.”
Will AI take over human jobs in hospitality? Unlikely, the panelists all agreed. It will, however, remove some of their responsibilities, freeing them up to do other tasks. At the luxury level, it’s about choice, Steele said. “Some people like tech; others like to have people. It’s about giving choice. Where we will see more technology is in the lower end segments.”
Years ago, the next big thing in technology were kiosks to check guests in. They didn’t have the best adoption rates and many hotels stopped using them altogether. Sage’s Schwartz called them the second stage of technology when it came to interfacing with guests, the first level a 100% human front desk. He said that there is a third level that is machine led and human assisted and a fourth stage that is fully autonomous. For instance, he believes that airport shuttle drivers will at some point go away because shuttles will drive autonomously. He’s not sold that the front desk will ever go that way. “If it is, we miss on the hospitality,” he said.