How artificial intelligence is thinking for your hotel

Artificial intelligence, still in its infancy, currently works best as an assistant. Photo credit: Getty Images/Pixtum

Bill Gates once said that we “always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next 10.” This is the kind of quote often shared before pitching an impossible technology pipe dream, but in the case of artificial intelligence the dream is becoming a reality. However, technology like this has no place in hospitality if it cannot be used to improve the guest experience.

Brian Kirkland, chief technology officer at Choice Hotels International, said his company is investing in AI to streamline operations such as improving purchasing decisions, or the timing of said purchases. He said back of house operations and even a hotel’s digital presence can benefit from machine learning, and in some surprising ways.

“Looking at how hotels and operators manage their business and compare that to data on how the industry and our competition is managing things, that is happening today. Having a system that makes decisions on your behalf can help you focus on more guest-facing programs,” Kirkland said.

FREE HOTEL MANAGEMENT NEWSLETTER

Like this story? Subscribe to Technology!

Hospitality professionals turn to Technology as their go-to news source for the latest technology products and trends. Sign up today to get news and updates on security systems, in-room entertainment, and more delivered to your inbox and read on the go.

True artificial intelligence remains out of reach today, but predictive analytics and machine learning is already at work behind the scenes. Existing systems are better tuned to work with numbers and take the guesswork out of analyzing trends, which is why Ron Pohl, SVP and COO at Best Western Hotels & Resorts, said hoteliers are most interested in applying this technology to the world of revenue management. Successful revenue management is now a 24-hour job, and having a predictive AI keeping your pricing competitive, tracking historical data and noting trend lines to then make recommendations for the future is a powerful ally.

“We are sometimes slow to react. Even if you have a full-time revenue manager there is a lag time between when you look at data and when you can use it,” Pohl said. “We addressed that to some extent with our ‘Autopilot’ feature, which gives the system some authority to make rate adjustments and we’ve seen some phenomenal results just with that.” 

Currently, AI is being put to use in the field of revenue management, where its predictive skills and reaction times are unmatched. Photo credit: Getty Images/Pinkypills 

Half a Mind

Finding a desirable guest-facing application for artificial intelligence in hospitality has been an enduring challenge. Many companies are playing with the concept of chatbots, which use predictive and analytics tools to respond to travelers, answering questions or providing suggestions when prompted.

On the surface this concept seems appealing, but Stephen Pratt, CEO at artificial intelligence developer Noodle.ai, said it might not be the best use case for the technology.

“Looking at all the places you can apply AI to your business, chatbots is one of the worst places to start,” Pratt said. “It’s very complicated to do technically, expensive and slow, and generally they don’t work very well. I would hate to see hotels start their AI journeys there.”

According to Pratt, AI is best used for solving problems beyond the complexity of what the human mind can compute. This makes them useful for large organizations, and not well suited for one-hotel operations, and they excel at predictive analytics.

“Eventually, hotel guests will expect the kind of predictions that only AI can achieve,” Pratt said. “We will start seeing a lot more targeted promotions and pricing soon, anticipating customer needs at scale. Once a hotel company completes the process, the dominos will start falling.”

Asset, Not Replacement

Because AI is so effective at what it does, fears abound about its potential to replace human workers, particularly in hospitality. Adams said this is not Best Western’s intention, but he admits he can see where these fears stem from.

“AI and machine language will not be in our industry to replace humans,” he said. “We recently traveled to a hotel in Hong Kong that used technology to augment its operations, freeing up time for employees to focus on what was important to customers. AI takes repeatable actions off of employees’ plates so hotel staff can spend more time with guests.”

Ramsey added that AI is also being implemented to improve hotel staffing, as machines are able to track trends and react to changes in events faster than humans, instantly responding to additional staffing requirements.

“Sure, AI works for operational scheduling, but can also make recommendations on how to solve problems,” Ramsey said. “It could help guests find answers to questions about their reservation, or can also work with staff to smooth operations. There are many opportunities, and they don’t eliminate jobs.”

Sometimes the illusion of AI will do just fine, even if it isn’t actually present. Ted Helvey, CEO and founder of voice and touch interface developer Angie Hospitality, said hotel guests want a personal, customized experience, but he admitted this can be difficult to provide consistently with current technology. Helvey’s company develops a personal assistant device that responds to voice and touch interactions, and draws on guest data retained by the company to tailor its responses.

The result, while not a true AI, Angie can work as a stand in.

“Angie can take room service orders, but we want it to ask guests ‘hey, do you want me to order your usual?’ We can keep as much data as the guest and hotelier want, and with enough these types of interactions become possible,” Helvey said.

Chatbots work as a form of machine learning, but are not true AI. Photo credit: Getty/Anyaberkut

On Guests’ Terms

Controlling your guest’s data—and reassuring these guests that, yes, their data is under control—is part of the technology race for two reasons. One, if your primary means of acquiring guest data is shared with a third party, you risk losing the exclusivity of that data. Two, guests want to interact with your hotel on their own terms, not yours.

Furthermore, with the European Union’s passing of the General Data Protection Regulation, all businesses will have to be more careful with consumer information going forward. This law addresses any data that is exported outside of the EU, and affects all organizations that do business with EU citizens.

“We are data rich in our loyalty program, and we know our customer there,” said Greg Adams, SVP and chief digital officer at Best Western Hotels & Resorts. “From a brand perspective, trust becomes a very important component of where your customers are going to spend their money. You can create all the algorithms you want to improve a guest’s experience… but a guest can now choose to eliminate all the data that makes this relevant.”

Another challenge hoteliers face in implementing AI is the legion of legacy systems still propping up many hotels from behind the scenes. Bill Ramsey, senior director of mobile and emerging channels at Choice Hotels International, said the advantages of having a modernized “stack”—otherwise known as the combination of software products and programming languages used to create a web or mobile application—are many.

“While working with legacy systems we are often forced to make concessions,” Ramsey said. “One workaround is to implement AI in silos. For example, one AI for customer care, another for loyalty. If you have a modern stack you can discuss the concept of a ‘single brain,’ but it will be a hindrance for a lot of folks.”

AI may be navigating a confused infancy, but its in hotel companies’ best interest to invest now. Adams and Kirkland were deliberately vague in discussing Best Western’s and Choice’s current AI projects, and with good reason. Right now it’s a confusing race, but still a race nonetheless.

“The cost of being late will be that your learning algorithms will have learned for a shorter period than others,” Pratt said. “The real advantage goes to those who go first. If you are late to the game, it could be the equivalent of a grad student competing against a high school student.”