The artificial intelligence chatbot tsunami has swept across all industries, travel being no exception.
But a recurring argument to temper the excitement in travel has been the fact that the underlying data of the large language models (LLM) used to train these bots is not providing real time data that could be used for transactional purposes.
Meet ChatGPT plugins. The latest initiative of OpenAI to cement its leadership among AI chatbots solves precisely this data dilemma. Companies joining the plugin program will provide the LLM explicit access to their proprietary, fresh data, allowing for training the model and enriching the answers with real-time data.
Inspiring and planning with AI - a match made in heaven
Since its launch last November, ChatGPT has already emerged as a disruptive force for the early phases of the travel customer journey. During the inspiration phase, these bots can provide targeted recommendations based on users' questions. For travel planning, it delivers relevant travel information with a few prompts, saving hours of online surfing across travel websites.
The industry has been quick in spotting the disruptive force of AI for these cases, with new travel startups popping up every day promising to deliver a better inspiration and planning experience. But even more relevant, some of the travel behemoths like Expedia and Trip.com decided to jump early on the AI bandwagon, releasing AI bots in their apps for trip ideas and planning purposes.
AI enabling conversational commerce
Tiptoeing AI for the early phases in travel is a newsworthy fact, but the industry has learned over the last decades that it is unable to compete against digital gatekeepers like Google to extract significant value out of this traffic.
This explains why the winners in online travel focused early on in scaling and optimizing their search and booking funnel, which led to a handful of travel giants ruling the travel space today.
Can ChatGPT plugins, capable of aggregating real-time inventory and pricing data, potentially trigger a paradigm shift in the current search and booking experience?
Subscribe to our newsletter below
Asking GPT-4 for the potential benefits and shortcomings of chatbot plugins versus current interfaces provides a fairly generic answer, mainly focused around the potential data accessibility constraints. The answer contains, though, a few valid points around bots' ability to:
- Provide a more personalized experience based on the conversational history
- Seamlessly blend a “human powered” support into the experience if needed
- Easily shift from written to voice interactions
The end of search as we know it?
The travel search result pages for flights, accommodation or rental cars have remained pretty much unchanged since the early days of online travel. Hick's famous law that the more options available, the longer the user will take to decide has not translated into significant improvements over the last decades other than adding a few more filters. Finding a better search experience to manage the trade-offs between comprehensiveness and cognitive overload has proven to be a hard nut to crack.
Travel search in a chatbot environment could certainly transform the endless list of search results into a few digestible answer nuggets that could be refined over conversational exchanges.
Considering how messaging has become one of the most prominent use cases of our digital lives, it does not sound crazy to assume that customers would be willing to engage into a conversational dynamic to select their travel options as long as a bot can provide fast and relevant answers, similar to what Expedia showcases in its plugin demo.
Plugins becoming the apps of the chatbot ecosystem
In the tech sphere, the rise of ChatGPT is sometimes compared to the pivotal moment of the iPhone launch. Could the creation of a plugins ecosystem have a similar exponential impact as the launch of the iOS App Store, sparking an explosion of new use cases powered by third-party companies and developers around the world?
OpenAI released 70 plugins in beta mode May 12, including Expedia, Kayak and Booking Holdings’ restaurant booking platform OpenTable.
Google announced May 10 that extensions will be part of Bard’s future ecosystem, showcasing Google Maps integration in a conversation with the simple prompt “show me on a map.”
Predicting how disruptive technologies unfold is closer to casino gambling than to science. Remember Steve Jobs’ famous first reaction to the app store concept was a four-word sentence: “It will never work!”
So some travel companies might be tempted to stay at the sidelines and watch how the current AI frenzy plays out. They better think twice.
Tech giants are racing to expand their AI tentacles across their entire ecosystem, including search and productivity tools. The underlying large language models are constantly being refined with trillions of datapoints at breakneck speed.
It will ultimately come down to user adoption if chatbot plugins will infuse a new conversational layer to the decade-old search and booking interface in travel. But travel companies that catch the AI wave early by starting to test chatbot use cases across their customer touch points today will be in better shape to survive the potential AI tsunami of tomorrow.
About the author ...
Mario Gavira is vice president of growth at Kiwi.com and an angel investor.