Using Artificial Intelligence to Predict Flight Delays

pexels-photo-723240

Flight delays are not just inconvenient for customers, they also cost airlines a pretty penny.

According to the book “Beyond Airline Disruptions” by independent airlines consultant Jasenka Rapajic, flight delays cost US airlines USD 22 billion annually. Another study sponsored by Federal Aviation Administration Air Traffic Organization Strategy and Performance Business Unit in 2010, puts this figure at a staggering USD 32.9 billion.

And this cost is only bound to go up, as regulatory bodies all over the world are pushing for compensation on behalf of passengers who are not forewarned of delays at least 24 hours ahead of the flights.

Loss of revenue from delays is only the tip of the iceberg, as passengers are prone to avoid airlines that are prone to delays. An independent study on the Thai low-cost airlines industry revealed that customer satisfaction and loyalty towards an airline has strong correlation with flight delays. As in, customer loyalty can suffer a serious blow if flight delays become a recurring thing, even in the low-cost airlines segment.

So how can the airlines industry tackle this sinkhole for revenue, and erosion of customer loyalty?

Thankfully, there is a solution. Artificial Intelligence and machine learning can go a long way in boosting airline profits by preventing loss due to delay.

Unlike popular culture and Hollywood movies would have us believe, AI is not actually omnipotent. So certain delays cannot be prevented by AI and machine learning alone. However, they can be predicted well in advance, provided that the airlines have rich historical data on flight arrival and delay times.

A highly stripped-down and bare-skeletal version of a dataset required by an AI to make accurate predictions would look something like this. The link points towards a dataset on flight delays and cancellations in 2015, created by the US Department of Transportation. The dataset contains essential fields like arrival and departure time, flight number, name of airline and the date. If we add additional variables into the mix like weather conditions, the pilot’s name and his historical records, details on the maintenance staff and their historical records, and any other factor that may affect delays like origin and destination airports, day of the week, or even the number of special-needs passengers on-board.

Even though AI cannot prevent flight delays altogether, at least it can help airlines prepare in advance and the customers can be forewarned so they don’t feel disgruntled and uncared for. Also, airlines can pinpoint areas of improvement where the lowest amount of effort will result in high, positive impact. Or send an automated warning to VIP frequent-flyer customers that a particular flight may be prone to delays due to weather conditions. That way, customers who are regular flyers and bring in high volumes of revenue, will feel like they’re cared for.

Thankfully, in the Southeast Asian region, namely Malaysia, there are AI and machine-learning solutions providers who have developed the capability to predict delays, with accuracy levels of up to 93%. And things can only get better for AI developers in Malaysia, as the Malaysian government introduces a USD 721 million fund for accelerating industries embracing smart technologies, such as robotics and artificial intelligence.

While the application for AI in predicting flight delays has the potential to save airlines billions of dollars annually, there are other areas where substantial impact can be made. For example, dynamic segmentation of customers into categories such as: high-profit, median-profit and low-profit; and offering each segment appropriate and automated incentives to cement their loyalty, can go a long way towards building a loyal customer base.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s