You’ve probably noticed how complicated it is to buy travel online when compared to other products and services. The reason is the large number of product variations. According to research Amazon sells about 700 billion different products, but Expedia customers have 1.26 quintillion (24 zeros in a row) options to choose from, what they’re looking for.
In other words, the search engine probably hits its biggest limit when using it to select one option from a truly vast range of possibilities.
Luckily, this is the kind of problem that technologies like ChatGPT, a generative artificial intelligence (AI) chatbot built on a large scale language model (LLM), are perfect for solving. Rather than search engines, users are often presented with an overwhelming number of results, often requiring filtering to find the best option.
One way to visualize the difference is to think of it as having two professional helpers. One of them, she has 10 possible answers to any question and can decide which one she thinks is correct. Already she’s the one who gets straight to the point and tells us what we need to know.
So it’s interesting that Expedia Group became one of the first major e-commerce operators to incorporate ChatGPT into their apps. Users of this feature (currently in beta for iOS users) can now plan their travel arrangements by having a free-to-interact “chat” with the application. This means discussing travel plans, the best accommodation options, sights to see and activities to experience, as if you were planning a trip with a human travel advisor.
How Expedia is using AI to simplify travel planning
Generative AI made headlines at the public launch of ChatGPT late last year. ChatGPT is a chatbot built on GPT-3 LLM created by Open AI. Of course, AI chatbots have existed before. Amazon’s Alexa and Apple’s Siri both fall into this category. But ChatGPT surprised users by allowing them to have far more sophisticated conversations. It can also generate texts such as articles, stories, scripts, speeches, poems, and even computer code to human standards.
Because this technology is accessible via APIs, many service and application developers are now racing to be the first to incorporate this technology into their services. For example, Microsoft (a major investor in OpenAI) added the feature to its own Bing search engine, as well as its recently announced social messaging app, Snapchat. Proprietary GPT integration.
They are all based on the same language model developed by OpenAI, but developers can add their own features to ChatGPT deployments. Bing powers the web search and enriches the training data we use to answer your queries with more up-to-date information. Snapchat integration also allows us to learn about you and grow your personality over multiple conversations.
Expedia’s innovation has made it possible to create a convenient list of options that are discussed during conversations while traveling. Users can refer to these to get an overview of everything discussed. By comparison, Vanilla ChatGPT tends to forget what was said between sessions.
Expedia also plans to notify users when there are special offers and discounts available directly through the ChatGPT interface.
Beyond this latest introduction of generative AI, Expedia uses AI in a variety of ways across its services. This includes providing personalized recommendations and results via search engines.
We also analyze historical trends in airline ticket pricing to track flight price fluctuations and predict the best time to book.
What are the challenges of using generative AI for travel planning?
One of the emerging limitations of generative AI chatbots is that they are prone to making mistakes.
Especially if you’ve used ChatGPT for some time, you’re aware that it can be misleading quite often. This trend, which affects not only OpenAI but also his other LLMs, is called “hallucination” because it shows that ChatGPT is sometimes out of touch with reality.
Expedia did not provide specific details about what steps it took to address the issue. It has been found that they may provide inaccurate or inaccurate information regarding travel bookings, such as stating that an accommodation or venue is wheelchair accessible or dog friendly when in fact it is not. I can’t help but feel that this could cause problems.
However, “Steps have been taken to limit inaccurate results and inappropriate responses. In some cases, the experience may not work as expected.”
Only time will tell how effective Experia Group has been in mitigating these problems, and reports of their successes (or failures) will soon emerge.
How might generated AI chatbots be used by Expedia and other e-commerce platforms in the future?
Expedia Group may be the first major e-commerce merchant to incorporate a generative chatbot into its core platform, but it is highly unlikely to be the last.
There are many ways you can use this technology to streamline your operations and improve your customer experience. These include:
An AI personal shopper that can analyze and interact with customer behavior and preferences to provide personalized service.
You can provide enhanced customer support that can more accurately answer customer questions and complaints and handle day-to-day tasks such as managing returns.
You can also create personalized product descriptions, so shoppers who arrive on your site see your product or service described in a way that shows how it’s specifically relevant to them.
This can be used to create interactive marketing materials such as personalized marketing emails and social media ads that allow customers to have conversations and answer questions.
It can also conduct market research, collect data from the customers they interact with, and provide natural language interpretations of behavioral data that give businesses valuable insights.
Whatever useful and innovative applications of generative AI emerge in e-commerce in the near future, it is clear that this technology has great potential to improve the user experience.
Like most people, I appreciate convenience, which is why I buy more goods and services online. On the other hand, it was sometimes frustrating because the options and outcomes were presented in an overwhelming way (this is especially true for travel) and it was difficult to find answers to certain questions. If generative AI can help solve these problems, it could be a game changer for the future of digital retail and commerce.