Understanding ChatGPT and its Implications for Hotels
You’ve seen the news. ChatGPT has been lauded and also scolded as a groundbreaking artificial intelligence chatbot that is the most adept to date at intimating human conversational versatility when delivering answers to any number of questions or creative requests. Compounding this is the news that Microsoft – an investor in OpenAI, the company behind ChatGPT – may be looking to incorporate this chatbot into its search engine, Bing, as a means of gaining market share versus the Google juggernaut.
Whether it is specifically ChatGPT or another service, it’s inevitable that advanced AI will soon be deployed into customer-facing entities like travel and hotel search. Hence, what should your hotel brand be aware of and how can you be ready?
A Cursory Understanding
To see how a chatbot like this can evolve hospitality web searches, you first have to know a bit – and only just a bit – about how the underlying software works. While admittedly we are not experts in the field of AI research, we emphasis for hoteliers that knowing how to use a piece of technology is far more important for our purposes than specializing in how it works.
For example, the majority of us know how to drive a car, and yet how many of us can fix one? That’s the mechanic’s job. It’s our job to source the mechanic we trust the most and pay them for their services. Apply this relationship to each hotel company and its respective vendors, tech or otherwise.
Nevertheless, there are some AI fundamentals that everyone should know. In this case, the acronym GPT stands for ‘Generative Pre-trained Transformer’ which means that the chatbot (in this case, GPT-3 or the third iteration) comes already trained using the full breadth of language-associated data on the internet as put through the cogs of 175 billion parameters in order to generate text-based information for its users.
The results are answers that are both highly realistic in terms of how an actual human would respond as well as highly accurate in terms of addressing what the user initially wanted. This GPT function is also reinforced with additional learning processes so that it gets more realistic and more accurate the more times it is used.
The Leisure Guest Right Now
To understand how this can specifically increase the utility and customer convenience of travel searches, let’s first examine the current behavior for how customers find then select a given hotel.
Currently, the holy grail of any brand is to be so top of mind that a guest directly accesses your website (or another owned channel) to complete a booking – no commissions and only first-party data. Think room blocks for a big event, corporate subscriptions, prearranged tours or browsing properties within a brand’s portfolio. Especially when it comes to FITs, though, this is often not the case.
Picture the average leisure guest looking for their next vacation. They spend time deciding their dates; they spend time in the ‘dream phase’ of choosing their destination; they spend time lining up flights or other means of travel that meet their needs. Frequently, they go back and forth on these values in order to optimize costs and any other form of agony. Hotel selections are probably downwind of all this, resulting in a host of customers who are relatively brand-agnostic yet have specific criteria already in mind – availability, location, budget, exclusive offers, room features, onsite amenities and so on.
The cross-comparisons herein to find the best fit can lead to dozens of browser tabs open at once, featuring a milieu of Google search, Google Maps, OTAs, Airbnb, TripAdvisor, brand.com, travel blogs and other resources. While this ‘sense of discovery’ is part of the fun, it can also induce ‘shopper’s paralysis’ – too much choice that sharply stresses the brain. If only there was a better way…
The Leisure Travel Opportunity
Given how versatile ChatGPT is as well as its capacity for further learning and refinement, such a vacationer can replace much of this helter-skelter travel search behavior with a simple chatbot conversation, akin to how numerous people today might call up their trusted travel advisor to outsource this decision-making mental fatigue.
You type in your search specifications and give some guidelines as to how deal-breaking each criterion is, then the AI returns with a narrow scope of recommended properties for you to further prod at with additional questions. (As a caveat, ChatGPT is still in beta but based on its preliminary outputs in the domain of creative writing, it isn’t a stretch to suppose that, in time, it will get good that interpreting your travel needs to deliver quite personalized hotel suggestions.)
Now let’s add some layers. Mix ChatGPT with ‘conversational AI’ language recognition and you have a very convenient and wholly automated form of voice-first hotel search – that is, no typing required. Next, throw in a virtual reality (VR) or mixed reality (MR) headset and you have a fantastic way to use emerging metaverse platforms to sift through immersive 3D representations of hotels or travel destinations, as guided by your voice-primary interactions with the chatbot software.
Content is Always King
Whether it’s Google, Microsoft, OpenAI or another leading tech company, most of this development is beyond any of our control. The best we can do is stay abreast with the latest news, comprehend the overall direction and then prepare accordingly.
In this case, the best way forward is to know that what’s always in your direct control is the ‘content’ about your brand that you then upload to allow these algorithms to crawl. We put the word content in parentheses for a reason here; particularly with the dawn of the metaverse, content itself is evolving.
For the past two decades, the meaning of web content has come to involve text descriptions, images, metadata, search engine optimization, paid search, user generated content, third-party reviews (and your responses to those reviews!), rate mapping to affiliates, social media, blogs, videos and so on.
Now, we all have to consider how our written words or imagery will be interpreted by AI – whether a text bot or a voice bot – so that our hotels are favorably presented, both in accordance with customers’ expectations about our properties as well as in adherence to given brand directives.
Lastly, hotels have to start thinking in 3D – recruiting vendors who can translate the physical site or individual objects into ‘digital twins’ so that these assets are ready to be deployed into a virtual world once the metaverse has (sooner than you think!) achieved a critical mass of consumers using their VR headsets to secure their next hotel reservation. The future is happening fast, so start planning now.