As CNBC reports, over 40 airlines have failed this year, and the outlook for 2021 isn't completely positive.

Although less than 50 cases of COVID-19 have been associated with air travel, people are afraid of getting on flights - and this fear translates to less revenue for airlines. United Airlines, for instance, expects Q4 revenues to be down a whopping 70% year-over-year.

Even before COVID-19, airlines have had a notoriously difficult time with customer satisfaction. To be fair, it's hard to keep customers happy when you pack them like sardines into a loud, flying tube. In the "new normal" of mask-wearing, distancing, PCR tests, temperature checks, and health forms, getting customer satisfaction right will prove even harder.

Airlines are now especially pressed for cash, which means they need to be picky about what they can improve - is it in-flight WiFi? Is it the meals? Is it leg room? Is it on-board entertainment? Is it the lounges? With the wide range of potential amenities to improve and attributes to tweak, machine learning is a great contender for the challenge. With AI, airlines can analyze their massive amounts of passenger data to figure out exactly how to optimize customer satisfaction.

How it Works

As with any tabular AI problem, optimizing airline customer satisfaction is a 3-step process.

We'll upload an airline customer satisfaction dataset to Obviously.AI, and select "satisfaction" as the column we want to predict. We can then see how each attribute impacts customer satisfaction, and deploy a model to predict customer satisfaction.

Insights

For instance, we see that business-class passengers have by far the highest satisfaction, but there isn't as an extreme difference between Eco and Eco Plus passengers. Armed with this insight, the airline in question might prioritize efforts to upgrade passengers to business class.

Given that most flights have empty business-class seats, a simple solution could be to offer extra seats at a discount.

We'll also see that in-flight WiFi is a major driver of customer satisfaction. WiFi is a relatively cheap expense for airlines, so they wouldn't want to cut corners here.

Naturally, seat comfort is another important factor, although it's not as important as in-flight WiFi or the class of the seat.

Every airline will have its own, unique set of customer satisfaction data. There are over 5,000 airlines in the world, so using AI to optimize customer satisfaction can give airlines a meaningful edge in this competitive environment.

In the same way, any business with tabular data - whether it's a SaaS company, a retail store, or even an insurer - can use no-code AI to optimize their KPIs.