Airline customer self-service with AI-powered conversation

You’re a business traveler whose plans have changed. You pull out your phone and …rebook yourself via chatbot?

To tackle this, we built an AI-powered, natural language engine that connects the British Airways NDC (New Distribution Capability) booking system. This proof of concept how natural language can be used by travel service providers such as airlines, travel management companies, airlines, and others with which a travel itinerary is present.

Using our API, the conversation engine can facilitate intelligent conversations on your website as well as popular apps such as Facebook Messenger, Slack, SMS. In the video we are connected to a mobile responsive chat window, which works equally well on a website or phone.

In the clip I requested flights from Boston to Manchester. Since there a two Manchesters with commercial air service the system asks the user to select the appropriate one. Next, I changed the destination to London Heathrow. This illustrates our system’s ability to respond to changes and understand various types of conversation input.

NDC sends back dozens of offers, which we grouped five at a time. To zero in on the flights most relevant to a user we added filtering. Asking for “morning nonstops only” reduced the list to two choices. It’s the British Airways NDC so the results are flights on BA or OneWorld carriers.

One differentiator between our approach and some of the other airline bots I have seen is that we’re keeping the user experience within the chatbot. This a challenge to execute gracefully with a transaction-capable chatbot but I believe it’s important for user experience continuity. Leaving the bot also sacrifices the valuable conversation data you’d otherwise collect.

AI-assisted customer service is when the bot hands off to a human agent. AI-assist is a trade-off between good customer service and the functional capability of a bot. Ultimately, the cost benefits, better user experience, and rich data of automated conversations will push aside the AI-assist model in favor of fully autonomous conversations; at least for non-status (Gold, Platinum, etc.) customers.

The first step in adding a chatbot to your digital customer experience  is to have a vision of what it is going to do for your users. Messaging and automation are already well entrenched your customers’ lives. When we asked business flyers where a chatbot might be invaluable we got plenty of suggestions.

With vision in hand your digital innovation team is equipped to evaluate the chatbot landscape for appropriate solutions to test. Those with solid technical capabilities might opt for an API-based solution that enables their dev teams to connect existing tools with an NLU engine such as ours. Tech-light companies could consider a white-lable solution that includes one or more customizable UIs.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.