Understanding GenAI with Its Applications that Hotels can Realize in 2024
In the first post of this series, we explored the value proposition of machine learning (ML) to the hospitality industry. It was a pretty straightforward article to write because ML is already generating demonstrable value for hotels and hotel companies, and there are still substantial financial gains to be realized by the technology. Now let’s talk about generative artificial intelligence (genAI), which many believe has the potential to dramatically increase productivity across all industries, including hospitality. How much of an increase? Estimates value the potential productivity boost to the global economy at up to $8 trillion a year.
As a result of that big number, there have been hundreds, maybe thousands, of stories written about genAI since OpenAI launched ChatGPT in November 2022. Many of them have focused on the long-term horizon of genAI – a lot of blue-sky visions. We’ll leave those posts for futurists and instead look at the short-term and mid-term potential of genAI for hospitality.
First, let’s talk about the power of genAI and what makes it so completely different from other forms of artificial intelligence, such as ML. In short, genAI can create new content – text, images, music, video, and even PowerPoint slides – by finding patterns and relationships between data sets of words, images, phrases, and concepts in large language models (LLMs), and using those patterns to generate something new. GenAI relies on ML models that mimic human neural networks, and it can produce a new output expressed in the form of a conversational experience that feels as if it was created by a human. This technology offers huge opportunities for productivity increases because its output is fully automated, and it can be applied to everything from repetitive accounting and employee support tasks to communication strategies that enhance guest satisfaction.
Let’s start with the guest. Enhancing guest satisfaction is a broad topic, so we’ll narrow it down to personalization and its impact on satisfaction.
Personalization has always been a goal of the hospitality industry, and to be clear, the ML that is already in place has been incredibly effective at creating predictions and prescriptions from large data sets that are specific to segmentation. Personalization results from an ongoing decision-making progression that evolves from making context-based decisions, or decisions based on broad market segment, to making micro-segmented decisions rooted in an individual’s real-time interaction through the reservation life cycle. Machine learning has given hotels and hospitality technology providers the ability to dig into the data and deliver segmentations down to these micro levels. In this instance, the machine learning model feeds off of the understanding of which offers the guest accepted versus which they ignored.
GenAI builds on ML, with LLMs populated by a range of data that can be analyzed for known and unknown relationships, creating new domain-specific content. That new content could include recommendations for hotels, restaurants, or local activities. It could also produce targeted content that details exclusive amenities and other services for one particular guest. For example, if a red-eye passenger is on her way, then GenAI could automatically offer early check in. Additionally, all of this new content can be delivered through any number of existing channels – email, web, apps, messaging, or even via an actual live person on the phone or at the front desk. The technology powers the interactions, which are then delivered in familiar ways to the guests.
With natural language processing (NLP), automated interactions can feel more personal and friendly, resulting in more satisfied guests. Historically, chatbots have been rules-based, with their responses keyed off by a specific word or phrase in the request – e.g., the guest asks a chatbot if the restaurant is open, and the response is a full list of opening hours of all outlets at the property. GenAI-based digital assistants can provide more specific and personalized answers. Instead of a list of mostly irrelevant facts, the response could be Yes, ABC restaurant is open tonight from 5pm-10pm. Should I book a table for you?
This example shows both the guest service benefit and the benefit of automation. The guest has the answer they want, and the automated service removes the need for an agent to answer that routine call and make a reservation.
Of course, this technology can be applied at the top of the funnel as well, providing personalized recommendations to the guest as they search for the perfect location or venue for their upcoming trip. Adding loyalty and guest history data into the model, then combining it with guest behavior during the search process can result in more targeted offerings, resulting in higher conversions and better utilization of facilities and services. A downstream impact to this type of conversational personalization in the booking channel could be a rise in attribute requests. After all, no leisure traveler is going to ask for a “double double;” they’re going to ask for two beds, and while they’re at it, they’ll probably ask for a view on a high floor because they can. Existing booking engines haven’t had a conversational component, but genAI can provide that function to make the booking process feel more like talking to a travel agent (remember them?).
Another use case for genAI is in content creation and marketing communications, including translations. Effectively using genAI in the marketing function can automate and amplify the reach of marketing and communication for a hotel. But how would a hotel get started? First, they should identify their need, which can be as familiar as a communication campaign to entice past guests back to the property during slow season.
When you provide genAI with a task, like a request to write a product description or an offer letter, it needs specific instructions and context direction. This could include product details, target audience, brand voice, and tone. The LLM processes this information and tries to understand intent. Then, based on its understanding of the instructions and the knowledge it has learned, the LLM starts generating text. It predicts the most likely word to follow in the existing sequence, and it even considers factors like grammar, style, and relevance to the request. This process continues until the desired length or format is reached. You can provide your own feedback to further tailor the result.
Finally, we can’t discuss genAI without discussing its potential for operational efficiency. We mentioned a few examples already, but here are some other functions that are already being addressed by software vendors in the market:
- GenAI can automate time-consuming tasks such as scheduling housekeeping, responding to guest inquiries, and managing reservations. This can free up staff time for other tasks and improve overall operational efficiency.
- GenAI can analyze data from hotel equipment to predict when maintenance is needed, preventing costly breakdowns and disruptions.
- GenAI can personalize room amenities based on guest preferences, such as adjusting the temperature, lighting, and music to their liking. This can create a more comfortable and enjoyable stay.
Obviously, genAI has the potential to significantly impact the hospitality industry in the short to medium term. By focusing on enhancing the guest experience and improving operational efficiency, hospitality businesses can leverage genAI to gain a competitive edge and create a more enjoyable and personalized experience for their guests.
Ready to geek out on how genAI works? Check out the genAI explainer piece here.
And stay tuned – our next piece will address how to find the right AI partner for you and your business.