We want to help brands in the travel industry stay informed about how they are considering and using generative artificial intelligence solutions from companies like OpenAI and Google. That’s why we’re surveying technology experts from leading travel brands and will publish the responses here regularly.
American Express GBT Chief Technology Officer David Thompson is the latest to provide insight into generative AI and its impact on the travel industry.
We started using generative AI. … We have been using AI across our portfolio of services for many years. Generative AI can be used to create different types of content, images, text, and videos. We use the open source Large Language Model (LLM) specifically to improve team productivity when considering applications that are used exclusively for text and provide written answers to questions. We are actively researching applications for
Our current generative AI efforts focus on: … applications with LLM have the potential to enhance rather than replace Amex GBT’s existing solutions. There are several potential use cases across internal functionality.
Our main focus at the moment is Traveler Care. LLMs can understand any amount of information presented in an unstructured or structured way. In seconds, you can answer questions in all areas of your booking and generate travel or approval emails at the same time. This allows travel counselors to complete many small tasks in a very short amount of time, reducing response time.
Travel policies are often unstructured and nuanced, but an LLM can help you read and interpret this information, allowing travel counselors to quickly answer your questions. For example, fare rules often consist of large chunks of unstructured text. The LLM helps travel counselors quickly understand it and see how it applies to the user’s travel policy.
Our biggest challenges related to generative AI are: …For us, one of the main challenges of generative AI is the prevalence of “model hallucinations”, the errors in the answers it provides. These errors undermine user confidence, so continually retraining the model to avoid them is costly. This is why we are looking at open-source LLMs that are not trained on sets of data, but work on vector databases. This almost eliminates the need to retrain the model.
Additionally, businesses, travelers and suppliers do not always have the same priorities and our industry is constantly changing. Managing the expectations of all these groups in real time requires an emotional intelligence that only humans can provide at this time. Innovative companies will experiment with long-context LLM with vector databases to explore the value they can add to managed travel while also recognizing the need to add an additional layer of human intelligence.
For the travel industry as a whole, we believe generative AI has the greatest potential. … Generally speaking, AI has great potential to personalize travel experiences and bring productivity to travel teams. LLM goes one step further and can learn from and make sense of large amounts of unstructured data, allowing applications to take over the repetitive and simple tasks that travel counselors need to perform, freeing them up to be creative. It can help you focus on problem solving and better solutions. Prioritize travelers.
A year from now, we expect generative AI to be used for: … We prioritize developments that increase the productivity of our traveler care team. LLM helps develop true co-pilots for travel counselors. Whether you’re managing recurring bookings or helping stressed travelers stay on track, it’s great for any task or situation. These developments will enable our Travel Counselors to manage large amounts of unstructured information and perform their duties quickly, freeing up time to better serve our clients and their travelers. Become. And as always, we will continue to ensure our solutions strike the right balance between technology, human interaction and intelligence.
