AI in Hospitality: Your AI may be ethical, but is it prudent? (Or is it just plain creepy…)
By Kelly McGuire and Arun Shastri
Arun recently posted an article on Forbes.com called: Your AI may be ethical but is it prudent?. The article dealt with ethics vs prudence and argued that something can be legal but not ethical and something can be ethical but not prudent. Which got us started thinking. Are there parallels in hospitality? We believe so. As hospitality organizations continue to explore personalization through predictive analytics and AI, it is becoming crucially important for leaders to understand where the lines are. What do you do when you’re feeling queasy about a new AI program, even if it’s arguably ethical? How do you decide whether to go forward? How do you know if you’ve crossed that line from providing excellent service to being a creepy stalker? A simple guideline discussed in the Forbes.com article was that the customer and the organization must both be better off after the transaction. In our case the guest and the hotel company. Let us consider a few cases in hospitality.
Consider an algorithm designed to use information about past stays to make promotions more relevant to a guest and tailor offers to guests’ behavior and preferences. This seems to be relatively innocuous application that should be beneficial to both guests and hotels. Guests receive more relevant communications and hotels should see uptick in conversions from campaigns. Let us extend this idea. What if the hotel wants to enrich their understanding of the guest by mining their social pages or credit card data? On the one hand, a bit more detail about guests’ preferences could dramatically improve the personalization of offers, but on the other hand, if guests know for sure they didn’t provide the information explicitly, it could start to feel like a privacy violation (stalker!) and be met with negative reactions.
Some revenue management thought leaders have been talking about customer-centric pricing as the next frontier for revenue generation for hotels. It is tempting to think that you could use consumer buying behavior to understand willingness to pay and value perceptions and then use that to create individualized pricing and offers. However, you may recall the extreme negative backlash against Orbitz when they simply changed sort order, not prices, because they found that Mac users tended to spend more on hotels than PC users. The same exact rooms at the same exact prices were there for both users, just presented in a different order, and customers still believed they were being taken advantage of. Incorporating willingness to pay into the promotion process, as some AI algorithms can, will surely improve revenue generating potential, but it is stepping into dangerous territory from a consumer fairness perception perspective?
Location, video and visual predictive analytics have a huge potential to improve guest experience and maximize operational efficiency. As technology has advanced, companies are now able to take advantage of location data, gathered via guest’s mobile devices. Delta recently unveiled some airport signage technology that will show personalized information to different passengers on the same screen at the same time, only viewable by each individual at the angle they’ve approached the screens. Here might be an example of value created on both sides, if the passenger gets valuable real-time information about their trip, privacy protected, while Delta has an opportunity to create a connection with the passenger and speak directly to them. If this were used strictly as a marketing opportunity, the passengers might feel differently about allowing Delta to use their location to speak to them.
Tracking how guests move through the property can help casinos put more profitable games in popular locations, or resorts place services or promotions in areas where they will attract the most (profitable) attention. The next step is to dynamically adjust on-property promotional signage or offers according to who might be in the property at any given time. For example, if there are families in the resort, you advertise kid’s activities. When there’s a corporate incentive group in house, you might be promoting a beverage program. That may be subtle, but the step after that is a “Minority Report” intervention targeting each individual as they walk by. This technology is possible but could be viewed as extremely invasive the more “personal” it gets.
Guests are likely aware of, and quite comfortable with, the notion that hotels have security cameras installed to protect their safety. This same video feed could be used to track and predict the size of the lines at front desk or restaurants, and proactively alert the need to add service capacity, or shift demand through off-period discounting. These use cases don’t track or recognize guests at an individual level, so relative privacy is maintained, and value is delivered on both sides. But what if an individual is identified? British Airways is using facial recognition to facilitate boarding at several of their gates. No need to scan a boarding pass, the gates open as the “recognized” passenger passes through. This potentially benefits both the passenger and the airline by removing friction from the process, but what if the same data was used to track movements of high-valued passengers through the airport to send location or behavior-based promotions during a layover, or even alert an agent to provide a personal greeting. Would this be considered too intrusive? Would guests feel like they were being stalked?
Facial recognition and visual analytics have advanced to the point that they can not only recognize a face, but also interpret emotions on that face. This might have a beneficial application in contactless interactions in the future, where the device could sense the mood of the traveler and potentially change the interaction cadence to improve the experience. Yet, wouldn’t it be a bit creepy if a robot said, “you seem upset”? Or if a manager approached a customer in the lobby from no where near the front desk and said “You seemed really pleased at check-in just now, I hope you enjoy the rest of your stay just as much”?
Your criteria should be that the hospitality company and the guest are both better off. Ask yourself, “Would I feel comfortable defending my program to a guest and making the case that it’s in their benefit?” Or would that be an uncomfortable exercise? There’s ethics and then there’s prudence, and then there’s just plain creepy.
