7 Ways Airlines Use Artificial Intelligence and Data Science to Improve Operations

Reading time: 14 minutes

Technology changes the way businesses interact with their customers, make business decisions, and build workflows. For instance, such actions as booking a flight via phone or conducting solely offline surveys seem unusual these days. Real-time access to data — the 21st-century oil — allows organizations to take informed steps towards operational efficiency.

We discussed with data science specialists and AI startupers how airline industry players use cognitive technologies to reach new heights.

The main applications of AI and data science in the airline industry

Artificial intelligence and its cognitive technologies that make a sense of data can streamline and automate analytics, machinery maintenance, customer service, as well as many other internal processes and tasks. So, AI technologies are useful for various aspects of airline operation management.

1. Revenue management

Revenue management (RM) is the application of data and analytics aimed at defining how to sell a product to those who need it, at a reasonable cost at the right time and using the right channel.

It’s based on the idea that customers perceive product value differently, so the price they are ready to pay for it depends on target groups they belong to and purchase time.

Revenue management specialists make good use of AI to define destinations and adjust prices for specific markets, find efficient distribution channels, and manage seats to keep the airline simultaneously competitive and customer-friendly.

Data scientist Konstantin Vandyshev, who worked at Transavia’s Revenue Management department, stresses that data science disciplines come in handy for achieving revenue management tasks.

Demanded flight routes. While RM is about finding the best way to sell a product or service, carriers use AI to answer one of the key questions: where to fly? “To define air routes, specialists have to analyze data and make decisions based on the insights. When researching a demand for a destination among different customer groups, they can rely on such data sources as search history and macroeconomic factors (e.g. GDP),” says Konstantin.

RM has industry-specific standards that specialists must use to define willingness to pay.

Willingness to pay. Collecting and crunching data about customers, airlines understand passengers’ tastes and behavior well enough to offer them transportation options they prefer and, more important, are ready to spend money on. So, revenue managers start from measuring willingness to pay (WTP). Willingness to pay reveals “when” a customer is likely to pay “a maximum price” for a product or service, explains the data scientist. “It’s assumed that customers are ready to pay more when there is less time before departure time. And society finds this pricing fair. WTP in the airline industry, therefore, depends on the day before departure (DBD). In practice, specialists define median WTP — a price that 50 percent of customers would like to pay for a ticket on a specific DBD. Such WTP is equivalent to price elasticity (the number of passengers that would buy a ticket if a price drops by a certain percent) with some assumptions between market demand and supply.”

This metric is connected to dynamic pricing — the practice of pricing a product based on a specific customer’s willingness to pay. The calculation of WTP requires selecting data correctly. Revenue management can combine similar markets and, alternatively, distinguish high and low seasons, as well as holidays and weekends.

“Approaches to this type of statistical analysis were developed nearly 10 years ago. These days, it’s easier to conduct research and present its results thanks to the development of data science and visualization capabilities. Considering that each case is unique, it’s very important to choose the right amount of data to extract insights from,” concludes Konstantin.

Expected marginal seat revenue (EMSR). This optimization model is calculated after WTP is defined. The metric can be perceived as the expected value of the current seat and entails allocating a seat to a specific fare class (FC). Data scientists measure EMSR by multiplying sales profit by the probability of selling an additional (marginal) seat belonging to the particular FC. “The moment comes when sales probability of a higher-fare ticket is so low that the expected revenue in a lower fare class will be bigger. So, knowing these probabilities you can determine the fare-class allocation for each day before departure,” adds Konstantin.

In a best-case scenario, specialists have to know the sell-up probabilities for different fare classes and days before departure to determine WTP and EMSR accurately, says this expert. The sell-up probability discloses whether a customer is likely to buy a higher-fare ticket if their request is denied. Clustering of flights according to destinations and flight dates is required. The revenue management team also carries out a clickstream analysis to know how many customers saw a web page showing a specific price. Airlines use historical sales data when determining willingness to pay and expected marginal seat revenue.

Ancillary price optimization. This is another approach designed to increase airline revenues through analytics-driven pricing. It allows data scientists to learn about a traveler’s tendency to buy ancillaries like baggage. Specialists define in which markets and on what days people are likely to pay more to check their bags. “For example, if I book tickets for three people with a child, then I’m ready to pay X euros more than if I flew alone somewhere on a weekend,” explains Konstantin Vandyshev.

revenue management in airlines

Revenue management starts with analyzing demanded flight routes

2. Air safety and airplane maintenance

Airlines literally bear high costs due to delays and cancellations that includes expenses on maintenance and compensations to travelers stuck in airports. With nearly 30 percent of the total delay time caused by unplanned maintenance, predictive analytics applied to fleet technical support is a reasonable solution.

Carriers deploy predictive maintenance solutions to better manage data from aircraft health monitoring sensors. Usually, these systems are compatible with both desktop and mobile devices, granting technicians access to real-time and historical data from any location. Knowing an aircraft’s current technical condition through alerts, notifications, and reports, employees can spot issues pointing at possible malfunction and replace parts proactively. Executives and team leads, in turn, can receive updates on maintenance operations, get data on tool and part inventory, and expenses via dashboards.

With applied predictive maintenance, an airline can reduce expenses connected with expedited transportation of parts, overtime compensation for crews, and unplanned maintenance. If a technical problem did occur, maintenance teams could react to it faster with workflow organization software.

Shane Ballman, former manager of Maintenance Systems & Technology at AirTran Airways and CEO of AI startup SynapseMX, Inc, came up with a platform that utilizes historical and real-time data to help maintenance crews make technical decisions faster.

We automate the routine and mundane using the company’s workflows, elevating things to the right person at the right time when a human’s touch is required,” says Shane.

The SynapseMX software analyzes data and metadata regarding detected maintenance activity. It helps engineers quickly evaluate a situation, for instance, to find out if this failure happened for the first time; if not, what can be done to fix it and how much time did it take to solve it previous times. Employees can also specify if there are spare parts available or a conflicting workload. “Then we evaluate for business rules — who cares, and under what conditions do they care? Should this trigger a new workflow? Update metrics? Send notifications?

Our AI is able to provide recommendations, in real-time, from the technicians in the field to the logistics team that supports them. The end result is a maintenance organization that reacts intelligently to current conditions,” concludes Ballman.

3. Feedback analysis

Air travel can be stressful even for frequent, experienced travelers whose passports are running out of clean pages. They have to do so many tasks like checking bags or finding a gate before getting themselves into a plane seat and taking a selfie!

In this regard, airlines that learn about pain points of airport and flight experience through data analysis can improve customer service. Using AI for feedback analysis and market research allows airlines to make informed decisions and meet customers’ expectations, agrees founder and CEO of PureStrategy Inc. Briana Brownell.

“AI systems can quickly allow airlines to determine if there is an opportunity to positively intervene in the customer journey and turn a poor experience into a delightful one.  It also allows companies to react faster in a synchronized, aligned way that is on-brand and consistent with the business’s values.”

PureStrategy introduces a platform for business analytics called the Automated Neural Intelligence Engine (ANIE). The engine’s functionality includes data review, categorization, visualization, and sentiment analysis. So, the engine does a lot of manual and time-consuming work with information allowing humans to concentrate on more complex tasks.

“We deal with customer feedback and voice of the customer data throughout all areas of the organization. Then, we link this data to internal operational metrics as well as external industry metrics,” specifies data scientist. Briana emphasizes the growing relevance of natural language understanding technology in processing and analysis of customer experience data as it allows for exploring the customer journey in their own words.

ANIE can be used to make sure whether it’s easy for customers to find, book, and pay for flights. “Ultimately we want to understand the ways in which an airline can delight a customer as well as where there is friction in the customer journey — and figure out how to fix it,” concludes Briana Brownell.

4. Messaging automation

When a disruption such as a flight delay or baggage loss occurs, travelers get nervous. And if customers don’t get a response or explanation of a problem from an airline representative in a timely manner, they likely won’t choose this airline for their next trip. The speed of response to customer queries matters as much as actual steps that are taken to solve an issue.

AI software, such as Coseer by Arbot Solutions, speeds up and simplifies customer service employees’ workflows by using algorithms for processing natural language or unstructured text. “We are helping airlines classify their customer emails and extract information from those emails so that they can automate some of the routine processes, for example, information about lost baggage,” says Coseer CEO Praful Krishna.

The solution can be used for chatbot development.

5. Crew management

Imagine a scheduling department that has to assign crews to each of thousands of flights operated every day. That’s a lot of work. Specialists consider a myriad of factors: flight route, crew member licensing and qualification, aircraft type and fuel usage, work regulations, vacations and days off to approve conflict-free schedules for pilots and flight attendants. Besides those are airplane maintenance schedules, training requirements like the pairing of senior crew members with junior ones, and government regulations that have to be taken into account.

“Crew management is a complex task due to many legal constraints. For instance, if staff belong to a trade union, limitations include an allowed number of flight hours and days off, as well as reimbursement in case of a labor law violation,” сlarifies data scientist Konstantin Vandyshev.

Of course, schedulers are not left to go it alone with big data generated by airlines (like, maintenance and passenger data, or data from onboard sensors.) Employees rely on software that integrates data from various sources, allowing them to get a full picture of daily operations. Using uncovered insights, they are able to make an optimal schedule in terms of working time, crew qualification, aircraft utilization, and expenses.

In other words, such software integrates predictive models with an airline operations management system.

Some crew management solutions allow addressing fatigue risk that pilots are in danger of due to a constant change of time zones, long duty days, scheduling changes, and other “pleasures” of working in the airline industry. For example, developers of the Crew Rostering solution from Jeppesen started integrating bio-mathematical models of fatigue into flight crew scheduling software. Their goal is to provide schedulers with the ability to rely on data about predicted fatigue to reduce risks in the planning phase.

At the end of the day, it’s all about safely transporting people from point A to point B.

6. Fuel efficiency optimization

Global aviation produces nearly 2 percent of anthropogenic carbon dioxide (CO2) emissions. That’s why aircraft manufacturers and carriers strive to improve their fuel efficiency. Well, it’s not only ecological but also financial concerns that drive airline industry players to use technology to reduce carbon emissions. According to IATA’s 2012 statistical compilation, airlines spend 33 percent of their operating costs on fuel.

Airlines use AI systems with built-in machine learning algorithms to collect and analyze flight data regarding each route distance and altitudes, aircraft type and weight, weather, etc. Based on findings from data, systems estimate the optimal amount of fuel needed for a flight.

7. In-flight sales and food supply

Eating a sandwich with a cup of coffee while enjoying the passing clouds and a bright blue sky is how many of us imagine an airplane journey.

Meanwhile, supply management specialists define how many snacks and drinks they must onboard without being wasteful. AI is here to help, too.

“If a low-cost airline sells food on board, it has to know how to predict the amount of food it has to buy for a specific flight,” says Konstantin. “While food isn’t expensive, every cargo load costs money. In addition, airlines usually throw away a lot of sandwiches at the end of a day. Companies that will be the first to solve this problem can make a good commercial use case out of it.”

Willing to dive deeper into the topic?

Ask your particular question at our Techtalks.

Explore Techtalks

Airlines using AI to improve operations

Delta: self-service for better customer experience and predictive aircraft maintenance

Delta, one of the world’s largest global airlines in the world, leverages AI to optimize operations and costs, as well as innovate customer service at every stage of a trip. The airline pays special attention to the airport experience. Almost a year after the introduction of a bag tracking technology, Delta presented a new enhancement.

Airport experience enhancement. In May 2017, it announced the $600,000-worth investment in four self-service bag checking machines. One kiosk was equipped to try face recognition technology to confirm a traveler’s identity by matching their faces with passport photos. Bag drops were installed in the Minneapolis-St. Paul International Airport. The airline stressed it will gather customer feedback regarding the new machinery to make sure it streamlines bag-checking. As of today, there is no information about the test results.

Airport experience enhancement

Delta passengers try out a self-service bag drop machine at Minneapolis-St. Paul International Airport. Photo source: startribune.com

Predictive maintenance. While an airline has no control over such disruptions as bad weather or air traffic control problems, it’s responsible for delays or cancellations due to technical issues. Delta is no stranger to predictive fleet maintenance. The airline has been using SmartSignal predictive analytics software since 2003, and integrated Bit Stew and Asset Performance Management (APM) both by General Electric (GE) in 2016. In short, these solutions filter and integrate data from physical assets, contextualize (add related information) it and provide actionable insights on their current technical condition. Guided by data this software provides, maintenance engineers work proactively and pull parts before they fail.

The asset performance monitoring program based on predictive analytics allowed Delta to improve such metrics as on-time performance. In the 12 months that ended March 31, the airline has managed to prevent 1,200 delays, service interruptions, and cancellations. In April, Delta said it plans to adjust the proactive maintenance program to newer aircraft such as the Airbus A350 and Bombardier CS100. With new models in the fleet, the airline will have much more real-time data for analysis.

easyJet: advanced analytics made easy

British low-cost carrier easyJet has turned operational challenges into successful AI use cases. In 2015, easyJet took a step toward becoming a data-driven organization when its former chief executive Carolyn McCall appointed the first head of data science department, Alberto Rey-Villaverde. The data science team, which keeps growing from its current medium size, included 25 specialists at that time.

The team collects and analyzes data on engineering, operations, market trends, and customer preferences. And now imagine how much data the airline that served more than 80 million people in 2017 generates. Exploring this big, heterogeneous amount of data facilitates solving diverse problems and coordinating numerous processes.

Airplane food supply. The airline has been using AI to determine how many bacon baguettes is enough to feed passengers on a single flight. The demand for this snack depends on such attributes as weather, types of passengers expected to be on board, and the time of the year. So, we can suppose that a predictive model included these variables among others. “We wanted to know the optimal amount of product required to meet demand without waste, so we are using machine learning to optimize how we load planes,” told Alberto Rey-Villaverde in the interview to eMarketer in 2016.

Revenue management. According to numerous articles, easyJet used data science to improve its pricing strategy and manage inventory. This approach allowed easyJet to increase profits per seat almost 20 percent between 2010 and 2014. The carrier also planned to analyze more than 1.3 billion searches on its site every year to determine optimal routes and flight times.

easyjet in the air - Revenue management

easyJet airplane in the air. Source: mediacentre.easyjet.com

Aircraft maintenance. easyJet continues incorporating AI in its operations. In March, the carrier announced its cooperation with Airbus under a five-year predictive maintenance partnership program. The airline will use the Skywise digital aviation data platform for predictive maintenance. With the platform capable of collecting 60 times more data than legacy systems, engineers will be able to replace aircraft components before they fail. That’s the way easyJet plans to reduce delays and cancellations caused by technical problems. The carrier says the new equipment will be installed on planes by next summer.

The airline also uses a recognition tool that speeds up passenger information processing. It reads the numbers from a document and fills out the information at the airport, so the traveler doesn’t have to type anything.

Southwest Airlines: operations optimization and excellent customer experience with big data analysis

Dallas-based Southwest Airlines is the largest low-cost carrier in the world and the largest domestic airline in the US. The airline calls itself a “Customer Service company that happens to fly airplanes,” and it seems that this strategy is working. Recently, Southwest was ranked the  top airline in North America and the sixth best airline in the world according to the TripAdvisor’s 2018 Travelers Choice Awards. Reviews collected from February 2017 to February 2018 were used for ranking; travelers evaluated airlines using such criteria as customer service, seat comfort, cleanliness, etc.

Great flight experience is the evidence that customer service quality meets traveler expectations. To maintain the fleet in good technical condition, control the work of each department, and know how a traveler feels about their trip, airlines strive to build their own data strategy. Southwest Airlines began collecting and analyzing their business data even before the big data hype began.

Workforce optimization. In August 2013, Southwest announced its intention to implement a suite of customer contact and workforce optimization software by Aspect, the airline’s legacy vendor of its automatic call distribution (ACD) system since 2001. The software suite includes six products that facilitate the interaction between contact center agents and travelers as well as optimize employee workflow.

Performance Management, for instance, provides front-line personnel with KPI dashboards on operational and strategic goals. With Interactive Tiles, contact center agents can monitor relevant performance metrics, access such data as daily tasks or schedules to stay productive. Another solution that helps understand customer intent and experience, Speech Analytics extracts meaningful information from recorded voice interactions between customers and airline workers.

Social media analysis. Listening to what your customers say and improving operations based on their suggestions are fundamental steps to achieving high-level customer service. That’s exactly what Southwest does in its Listening Center that was opened in 2014. Forty experts from Customer Relations, Marketing, and Communication departments monitor social media feeds to allow the airline to solve emerging issues as rapidly as possible. People actively share their travel experience on social media these days, so a department like Listening Center provides invaluable input for operations management.

Team members track sentiment on social media about Southwest, its competitors and the airline industry as a whole, analyze the most popular industry-related topics discussed on social platforms, and follow news from traditional media. They also answer travelers’ questions and respond to posts mentioning Southwest on social nets. And they do this day and night. Insights from social media help the airline stay current with trends and operate efficiently. Besides this, analysis of real-time data from social media allows Southwest to provide customers with personalized offers.

Southwest listening center - Social media analysis

The Listening Center specialists explore data from various resources 24/7. Photo source: Southwest Airlines

Air safety. The airline also puts big data to work to fly more safely. Southwest has been working on a text data-mining project with NASA since 2008. The specialists use algorithms to text-mine large amounts of data generated by airplanes, as well as report data from pilots and others like air traffic controllers, to detect patterns aiming at potential safety problems. Sadly, incidents leading to lethal outcomes may happen even with airlines known for their obsession with safety. April 17, an airplane made an emergency landing at a Philadelphia airport after one of its fan blades had broken off due to metal fatigue. Since it was a jet engine that exploded, the manufacturer should take a responsibility for the failure.

Fuel consumption optimization. In 2015, Southwest signed a contract with GE Aviation to use its flight analytics system to improve fuel consumption for its fleet of more than 700 Boeing 737s. The cloud-based system that runs on the Industrial Internet allows for collecting and analyzing data generated by aircraft during a flight. For example, pilots can consider the information about wind speed, air humidity, plane weight and speed, maximum thrust, and altitude when planning the amount of fuel needed for the next flight to the same destination.

The future of AI in the airline industry

Today, AI makes it possible to enhance customer experience with automation and self-service solutions, optimize employee workflow, and ensure higher air safety with predictive and prescriptive aircraft maintenance. It also allows airlines to make informed decisions about pricing and market positioning through the smart use of data.

“It’s worth mentioning various stochastic optimization tasks in regards to potential use cases of AI in the airline industry. Eventually, data science is applied to operations optimization, but carries great hope for the new technologies development,” notes Konstantin Vandyshev.

Briana Brownell also admits AI’s key role for operations optimization. “I see many opportunities! For instance, to optimize operations including adding, changing or removing routes, setting flight times, pricing, and product offerings. Ultimately, success is driven by having a deep understanding of different customer segments and where new market opportunities exist,” concludes this data scientist.

We’ve told you about a few airlines and their data science initiatives. Feel free to comment how your business uses AI in the comment section below.



Sort by: newest | oldest | most voted
Jul 12, 2018

So now are we using the term “AI” so loosely to describe anything related to data science? The article claims that AI is used in activities such as Revenue management and Predictive Maintenance but what it is describing is just the use of data science, which is nothing new.
Articles such as this one are the reason why this field is so full of hype.

Jul 19, 2018
AltexSoft Team

Hi! Often, we combine these concepts for educational purposes and to make our stories accessible to a wider audience. We understand the confusion and have updated the title accordingly.

Mar 20, 2019

Artificial Intelligence in Aviation Market worth USD 2,222.5 Million by 2025 https://www.marketsandmarkets.com/Market-Reports/artificial-intelligence-in-aviation-market-106037016.html

Mar 21, 2019
AltexSoft Team

Prashant, thanks for sharing the information!