conversational AI, airline, aviation

Sometimes the best learning experiences in business grow out of the worst snafus. But often we take away the wrong lessons from a business disaster.

Such was the case with recent stories in the media about passenger reactions to automation run amok: Frontier Airlines turned off its telephone help lines, replaced them with chatbots in text messaging and enraged its customers. The lesson that the media reported? Automating telephone help lines with texts, generative AI and other technologies represents a defeat for consumers and a victory for money-hungry corporate managers.    

I couldn’t disagree more. To me, it was as if the media covered the sinking of the Titanic and concluded that it was time to stop building cruise ships – rather than build cruise ships better. The lesson of the Frontier Airlines kerfuffle is not that automation of customer service inquiries using generative AI is a bad idea. It is that the execution of the idea in this case was terrible.

When Frontier told customers last November that it had terminated all customer service by telephone, what followed was a wave of negative news stories about a budget airline, its attempt to cut costs and the frustrated, angry customers who could not get the information or help they needed. But the story could have been different. 

When done correctly, the addition of text messaging and the automation of inquiries with conversational AI can be a victory for consumers, companies and the many worn-down customer service agents who handle the calls. When it works right, text messaging and bots give consumers communications channels they prefer, give companies greater efficiencies and improve the lot of agents who experience less of the grueling, monotonous work of helping (often angry) customers on the phone. 

AWS

First, for many people, telephone calls are not the preferred means of communication. Younger consumers often reject telephone calls outright – when was the last time you spoke with a Gen Z on the phone? And for passengers in transit there may not be time for a phone call while rushing to try to catch a flight. 

When one of our clients gave customers the option of switching from telephone communications to text messaging, we saw customer satisfaction jump by 14% for the texting customers. (Note that I said we gave customers the option; we didn’t force them to use text messaging.)   

In addition, texting and other electronic forms of communication are much better to automate. An airline can transmit flight schedules in table format and images of maps that can tell you exactly where a gate is located. It is easier to get more information across faster in text than in voice.   

Finally, you can make your customer service agents happier employees. Agents for an airline must contend with a stream of unhappy, sometimes furious customers hour after hour, day after day. The work is so emotionally draining that agent turnover is high. Beyond the human toll, that’s a big cost to companies to recruit and train every new employee. 

By contrast, an agent working with customers on text messaging experiences far less stress. The psychological impact on an agent from a “screaming” text message in all capital letters is far less than a screaming customer on the other side of the phone.

If an airline can convert the 10,000 telephone calls it receives a month to 5,000 telephonic and 5,000 digital inquiries – some answered by conversational AI and others by human agents who are more efficient – it can reduce the cost of its call centers by 20% or more.  

So, what should Frontier have done differently? When the airline announced its plans, it came as a sudden surprise to customers. Frontier should have slowly transitioned its passengers onto an excellent experience implemented on digital platforms while leaving customers who felt they had to speak with a human being the option to do so.

Most importantly, Frontier failed to deliver a great experience in their new digital channels. The chatbot Frontier used to replace human call centers was hopelessly basic. It failed to take advantage of conversational AI, which can respond to individual queries by understanding the intent of the language. 

The Frontier automated system gave customers static information from their website instead of providing actual answers. The system did not easily help passengers get answers because it was not integrated into Frontier’s reservation system. For example, to answer the query, “My flight to LAX was canceled, and I was rescheduled onto a flight I don’t like!” AI could respond with, “Here are some alternate flights to LAX you can choose…”. Instead, it responded with impersonal instructions about how to change the flight on the website.

The lesson of the Frontier debacle is that great execution in business is more important than good ideas. Frontier’s intent to serve customers via text messaging with conversational AI was good, but they failed on the execution.    

TECHSTRONG TV

Click full-screen to enable volume control
Watch latest episodes and shows

AI Data Infrastructure Field Day

TECHSTRONG AI PODCAST

SHARE THIS STORY