How to Maximize Generative AI’s Capabilities
Predicting in 2023 what generative AI (think ChatGPT) will mean for the hotel business is a bit like presaging in 1995 what the web would mean for the hotel industry. The web’s potential may have been clear enough back then, but Expedia and Booking.com wouldn’t launch until the following year.
The industry made that transition and, generally, gained from it. Hotels upgraded backend systems and developed the customer-facing interfaces that let us browse properties and book without making phone calls. Building on those foundations, along came Airbnb and VRBO.
Generative AI’s impacts, if yet unclear in their specifics, promise to be similarly profound. And just as was the case a generation ago, hotels will need to up their IT games to exploit the possibilities this nascent technology is poised to present. Hotel business leaders must identify fruitful AI applications while hotel IT departments must hire strategically and understand how to gather and structure dispersed data to maximize AI opportunities.
Those opportunities fall into two general categories. First is an ability to harness AI to personalize customer service, improve customer experience, and boost customer loyalty while reducing costs. Second is about developing new revenue streams. The basis for both will be to combine ChatGPT-style AI that uses immense training data sets as a foundation with a hotel’s own AI capabilities. That localization, which includes a range of guest- and business-related data, will improve model accuracy to the hotel’s great benefit.
Hit the Slopes
To explore AI’s potential on the customer front, let’s think about a hypothetical. It’s not hard to imagine layers of AI enabling highly curated experiences with minimal human intervention. Someone planning a ski vacation could reserve and pay for flights, ground transportation, hotel accommodations, rest-day tours, restaurant experiences, ski rentals, passes, and lessons and more with a few keystrokes and confirmation clicks (or, if via mobile, taps).
Hotel-specific AI will play an important role in all this. It will optimize room choice or even hotel choice based on guest profile (a couple versus a family with young children, say) and adjust for local variations and recent wrinkles that may elude a more generalized AI system – a restaurant with a new chef or a forecasted snowstorm that could render an off-day snowmobile tour too great an avalanche risk. This sort of personalized service will improve the customer experience, build customer loyalty, and, over time, grow revenue.
Key will be to own the guest experience and capture more wallet share before, during, and after the visit. By profiting from referrals to restaurants, ski resorts, and from tour and transportation companies, a hotel stands much to gain on the financial front. Getting there requires not only data-driven insights into what different sorts of customers are most likely to want, but also detailed tracking of both booking and, on the partner side, confirmation of service delivery.
Closing the Loop
For a hotel, a key side benefit of harnessing AI will be an ability to close what has been an open loop, where so often referrals are thrown over a figurative wall, with the benefits accruing solely to the business getting the referral. Hotels can also filter AI-derived recommendations to the advantage of preferred partners.
Closing that loop won’t demand the wholesale replacement of the potpourri of systems hotels currently use. But it will require new tools capable of collecting data from various silos and amassing them in the data warehouses that AI systems thrive.
In the ski-resort hotel scenario, that data will be diverse. It will include inputs from public sources such as weather, traffic, school-vacation, local-event, and other data; supplier and service-provider availability; customer-related data such as place of residence, age, income, social-media, and inputs from past customer feedback; and data from internal systems tracking website traffic, sales, guest information, room bookings, property management details, and billing information.
Data Plus People
The power of a hotel’s AI will improve over time as the data volume expands. A ski-resort hotel’s system may calculate the probability of a prospective guest needing to rent equipment – and whether that equipment suggestion should be for a high-end demo skis or standard rentals based on distance from the resort, travel mode, rental history, age, and other factors.
The impact generative AI will have on hotel staffing – one of the industry’s biggest challenges – remains to be seen. It could be that expensive properties featuring personalized, concierge-level service maintain the status quo. With AI, though, hotels at lower price points and fewer staff may be able to provide comparable offerings with lower labor costs. Given Millennial and Gen-Z preferences for planning and managing their lives on screens, not developing compelling AI solutions could end up being a dire threat to the business. Given the upside of AI in the hotel business, the industry’s leaders will need to do more than just invest in data warehouses and AI applications. They’ll also need to attract – or outsource – the IT talent needed to make this transition.
When Instagram was founded in 2010, who could have predicted that a photo-sharing app would bring about the redesign of hotel lobbies and influence so many influencer-inspired bookings? Similarly, just how generative AI will change the hotel business remains to be seen. What’s obvious already is that hotels have much to gain from it – and, conversely, they ignore it at their peril.
About the Author
David Rapp is a solution manager for travel and transportation at SAP.