Conversational AI in Travel & Hospitality

The art of being hospitable

Conversational AI, chatbots, and voice assistants, are widely used across the Travel and Hospitality industry. I had the opportunity to chat with a panel of experts to learn more about how brands are leveraging these solutions.

The panel included:

• Alan Kiernan, CTO, Parly.ai by Cation Consulting
• Andrei Papancea, CEO, Co-founder NLX.ai
• Laetitia Cailleteau, Managing Director, Global Lead for Conversational AI, Accenture
• Massimo “Max” Morin, Head of World Wide BD for Passenger Airlines, AWS

Watch the video and read highlights below

Why conversational AI?

Customer care and customer support are at the core of the Travel & Hospitality industry. The key piece is being hospitable — taking care of your customers and your business. Brands show value and differentiate themselves based on their level of customer care. For example, a high-end luxury brand will provide higher levels of concierge service.

Conversational AI, chatbots, and voice assistants are a key component of customer care. They help customers get the information they need or complete their goal, with a low level of friction, 24/7, on channels users prefer to interact.

Conversational AI enables brands to scale operations and adapt quickly, at low cost, in a practical way. For example, the airline industry has been hit particularly hard through the pandemic with lower volumes of passengers. On top of it, the world is unpredictable — events like a volcano erupting, flooding, or even the current war in Europe occur. People still need to travel during these times, and operations need to be stood up quickly. As Laetitia Cailleteau of Accenture explains, virtual assistants can be put in place quickly, and can be adapted and updated rapidly to handle new queries as a result of these events.

Alan Kiernan of Parly.ai seconds this ability of conversational AI to react quickly to change. When Covid-19 hit, airlines faced a lot of requests for refunds that put enormous pressure on call centers. Chatbots and intelligent Interactive Voice Response (IVR) systems enabled airlines to handle these requests in an automated way that was fast and scalable.

Conversational AI is not purely an automation play, as Laetitia points out. Call centers have been scripted to death. These scripts were system focused — written based on how to interact with the internal systems — rather than human focused, and have been written ages ago. The importance of chatbots is to spin the experience on its head, instead of automating the existing experience. Today we need to be human centered and think about the service we are providing for humans.

There is a balance to implementing conversational AI, as Andrei Papancea of NLX.ai indicates. One cannot achieve 100% automation, without annoying customers. There still needs to be a path to escalate when needed.

Handling complexity

While enterprises across industries tend to start with high volume, low complexity tasks, there are use cases in Travel and Hospitality that may appear relatively straight-forward, but are quite complex behind the scenes. These use cases, while harder to implement, given disparate data sources and complex rules, are worth doing.

For example, questions related to baggage handling, allowances, and tracking can be quite complex. As Laetitia explains, there are a lot of contracts between airlines with code-share agreements regarding baggage. A suitcase that may be acceptable by one airline on a code-share, may not be on another. The challenge for conversational AI is not in automating a solution to answer questions related to baggage, but in finding the correct answer. The first airline may say one thing, the second another, and the call center a third, so the contract between the airlines needs to be checked.

As Alan adds, conversational AI is a great example of a user experience that can be simple and direct, while dealing with a complex issue behind the scenes. For example, travel restrictions and rules related to Covid-19 within Europe differed widely across countries and changed frequently. Chatbots needed to be able to surface the data from reputable sources, in multiple languages, and to keep updating automatically for the rule changes — all seamless to the user.

Personalization is key

Context and personalization are key to a better customer experience. The bar in Travel and Hospitality is quite high. As Andrei explains, when a customer contacts a company, the company needs to know who they are and the context — whether they booked a flight or hotel already and why they might be calling.

When an IVR answers and greets the customer by name and recognizes their recent activity, it helps put the customer at ease, and comfortable interacting with the virtual assistant, indicates Andrei. It sets the tone of the conversation.

As Laetitia points out, the less you have to ask a user the better, as it means you know them well. While we are excited about Natural Language Processing (NLP), if you design the experience well, know the customer well, and understand what they may want, you can propose things to them with little input required and the result can be a quite cool, slick experience.

Multimodal choice

Conversational AI solutions are part of omni-channel, multi-modal customer engagement platforms.

As Max Morin of AWS points out, multimodal solutions enable brands to take full advantage of the device. If a customer is calling an airline to change their seat, the IVR can send a link to the website to show the seat map while helping guide the user in selecting a seat. Guests and travelers are becoming more accustomed to using technology when interacting with large tech firms like Amazon and Facebook and expect similar experiences with Travel and Hospitality companies.

Andrei adds, multimodal solutions can be very practical, as some use cases are impractical on certain channels. For example, capturing a customer’s name and booking reference number via voice IVR can be challenging. Instead, the IVR can branch out to an additional channel to help the user through the task. As Andrei recommends, leverage the power of each channel and help drive the user to the best possible experience.

Regardless of the channel, it is important to have one source of truth for the customer. In Travel and Hospitality, there are different customer journeys — pre travel, day of travel, post travel, and life of the customer, as Alan explains. Many companies and systems in the industry still rely on paper and PDFs. It is important to pull the data and join it to the life of the customer, as the information aids in personalization and can drive huge value.

As Laetitia adds, it is important to provide choice and flexibility. Use the right channel for the right task, but also be flexible to what the user wants. It goes back to being human-centered. We are all humans and may do things differently based on our preferences or even generational differences.

Human agents are important

Chatbots and virtual assistants are not meant to completely displace human agents. As Max indicates, conversational AI can take on the high volume, low complexity tasks, leaving human agents to do things only humans can do — handling the more complex issues. This helps reduce the stress on the agents, not having to answer the same question every time.

Chatbots and virtual assistants also enable more efficient routing — passing along the context of the interaction to help the agent resolve the customer issue.

Some companies pride themselves on containment, but at the expense of customer satisfaction and the quality of the experience. The hospitality component is lost at this moment, as Andrei points out. If someone wants to talk to an agent, take them to the agent. Do not try to fool the customer. If the virtual assistant can handle the task well, handle it, otherwise take them to an agent.

The power of words

As conversational AI experiences continue to improve and advance, there is more interest in fine tuning the conversation to be more empathetic.

How a chatbot or voice assistant phrases a question will have an impact on how the customer responds, which can impact whether the conversation goes down the right path or wrong path. It is important to be nice, approachable, sympathetic, and perhaps sometimes funny depending on the context. As Alan explains, the goal is to be Jarvis, not Terminator.

It is helpful to have short and concise conversation.

As Laetitia puts it, conversational AI is about the “power of words.” With web development, designers did not have to be as cautious on the length of content. With mobile apps, there was less real estate. However with conversational AI, it is even shorter. Users want a short, concise answer. It takes a particular skill to craft good dialogue.

Conversational design is an art form as Andrei adds — words matter. This is why we sometimes see former screenwriters as conversational designers as they have skills with words and dialogue.

Understanding the user, who they are, and the context, helps establish trust and empathy. When a user makes contact, the company knows the user is calling, they know there is a high propensity the user is having a problem, and they have context. Empathy is part of the handshake between a human and machine, as Laetitia points out. It is important to make sure you have a great design and a well trained NLP model to help build the relationship and trust.

Max adds that it is also important to understand what might cause a user to be unhappy or drive anxiety. For example, if a customer has to repeatedly identify themselves, or if a navigational menu has too many options to remember, or if the user has to repeat themselves multiple times, those can cause frustration. If the user is calling from their phone, the company should know who they are; and if they booked a flight or hotel, it should know that as well. The technology can leverage all the information it has about the user and context to help preempt customer frustration.

A hyper-personalized future

Our panelists all see the level of personalization in conversational AI going even further for Travel and Hospitality.

In the 1980s customers went to a travel agent for a personalized experience; in the 1990s, customers went to the Internet for self service; and now everything is a-la-carte unbundled, as Max explains consumer behavior over time. We are starting to see a desire to go back to personalized engagement.

Max is looking forward to the day that conversational AI can recommend the best travel options for him, and he can decide in confidence and comfort knowing that they are the best personalized options for him, based on the relationship built with the service and the information it knows.

Alan adds to this hyper-personalization with the hope of a concierge service that continues to evolve based on his activities. If he books a trip, the conversational AI concierge should notify him of activities and special offers he may be interested in, that it figured out for him, and he can just respond “do it.”

Andrei would also like to see this level of personalization unified, and carried over across experiences, transcending companies. We have preferences that we may have explicitly set or can be implied based on our behaviors. The goal is for the technology to “just know me” regardless of which company one is interacting with or channel interacting on. As Andrei indicates, people are willing to provide data if you are willing to provide something of value in return.

Laetitia also envisions a more personalized experience, at scale. In the past, it was not possible for a human agent to remember all the data about a person, or calculate a user’s propensity to do something so quickly when answering the phone. We are starting to be able to do that. Laetitia also sees further digitization of information and incorporating additional data to augment personalization. For example, if an activity is booked, but the weather is bad, have the conversational AI recommend an alternative based on preferences.

It is exciting to see all the advancements in conversational AI in the Travel and Hospitality industry. I look forward to seeing new use cases and further incorporation of personalization into the experiences.

Arte Merritt leads Conversational AI partner initiatives at AWS. He is a frequent author and speaker on conversational AI and data insights. He was the founder and CEO of the leading analytics platform for conversational AI, leading the company to 20,000 customers, 90B messages processed, and multiple acquisition offers. Arte is an MIT alum.