Ride The Wave

The Impact Of Automation On Airlines

By Ben Jenkins

For years, airlines have realized significant benefits from automation, and the use of automation will only increase in the years to come. While automation will impact nearly all job roles to varying degrees, it should not be feared. The rise of automation in airlines will unleash people to pursue more creative and higher-value endeavors.

Ascend Industry

There is no question a new wave of automation throughout the workplace is upon us, thanks to recent and continuing advances in technologies such as artificial intelligence and robotics. These technological advances will impact almost all jobs within the workplace today and may even eliminate a few altogether.

Airlines, too, will be affected by this new wave of automation and, in many respects, are riding the crest of the wave. Already at some airlines, according to Mary “Missy” Cummings, director of the Humans and Autonomy Lab at Duke University and former Naval fighter pilot, “pilots actively steer aircraft for just three to seven minutes of many flights, with autopilot guiding the rest of the journey.”

Recent research conducted by U.S. based McKinsey & Company suggests as much as 45 percent of the work humans are currently paid to perform can be automated using existing technologies. This is not to say this amount of jobs could be fully automated, but potentially up to 45 percent of the activities performed across all roles could be. About 60 percent of jobs could have 30 percent or more of their existing activities automated and 20 percent of roles could see 70 percent automation, representing a significant transformation of job roles in the future.

The obvious benefit for businesses, including airlines, is labor savings. Additionally, automation often results in output volume, speed and quality increases. Furthermore, the new wave will not just impact low-skilled and low-wage roles, it will impact nearly every type of role to a certain degree. To this point, the research currently suggests that more than 20 percent of existing work done by chief executive officers could be automated. The benefits also compound as employees are freed from the burden of monotonous manual work and are able to focus on high-value tasks. Employee satisfaction with their work environment would increase, leading to more productive employees who are focused on work that delivers a higher return. This includes activities that are more creative and emotional in nature ... in essence, those that require a human element and are far more di cult to automate.

While the benefits can be significant and wide ranging, there are downsides to automation. The more sophisticated the automation becomes, the less astute the human operator also becomes, who now only supervises what has been automated.

Moreover, like all technology, from time to time, automation fails. These failures tend to be acute and severe and can only be arrested by a human operator. Except the humans may no longer know how, having fallen out of practice.

This is known as the paradox of automation, and the airline industry knows its perils better than most.

Robotic Process

McKinsey & Company suggests data collection and data processing represents a third of all activities performed across job roles in the United States. According to the company, “both activities have a technical potential for automation exceeding 60 percent.” Considering the number of roles across an airline’s back office that have a high percentage of activities related to data collection and processing, the impact on airlines will be profound. This will not only include roles within traditional business functions such as finance, human resources and information technology, but it is equally applicable to all airline-specific functions.

An area where many airlines have benefited from automation for years is that of robotic process automation (RPA). “RPA is a type of software that mimics the activity of a human being in carrying out a task within a process,” said Leslie P. Willcocks, coauthor of ‘Service Automation, Robots and the Future of Work.’

Airlines have used the simplest form of RPA for some time, including ticketing robots and flight-firming robots, for example. There are, however, newer and more advanced forms of RPA only just starting to take off. These newer forms can execute in real-time, be based on multiple data sources and undertake much more complex decision-making algorithms. RPA can deliver return on investments up to 200 percent, although it is no match for machine learning and other forms of artificial intelligence that can continually learn and apply that knowledge to what they are tasked to do.

Machine learning is a branch of artificial intelligence linked to the development and application of adaptive algorithms that use example data or previous experience to solve a given problem. Such algorithms overcome following strictly static program instructions, as used in RPA, to make data-driven predictions or decisions.

The technique is being applied to revenue-management activities by some airlines. Through machine learning, a system learns information from pricing interventions a pricing analyst makes, based on the data at the time, and then takes over responsibility for making that same pricing intervention.

Beyond learning from humans, machine learning can then determine how successful its own intervention was compared to other interventions it made and improve future interventions. This may be the degree to which a price was adjusted up or down or whether the intervention should have been made at all.

ACCORDING TO MCKINSEY & COMPANY, UP TO 45 PERCENT OF THE WORK PEOPLE PERFORM TODAY CAN BE AUTOMATED USING EXISTING TECHNOLOGIES. THIS DOES NOT IMPLY THAT 45 PERCENT OF JOBS COULD BE REPLACED WITH TECHNOLOGY, BUT THE AMOUNT OF WORK BEING DONE MANUALLY COULD EASILY BE AUTOMATED.

Impacts On Airline Job Roles

One such area modern RPA can be applied is that of compliance. Airlines are heavily regulated, and many have entire departments devoted to ensuring they meet their regulatory obligations.

“Compliance is ripe for automation because it is both rule-based and data-intensive,” according to a Harvard Business Review article titled “The Knowledge Jobs Most Likely To Be Automated.” “The more rules there are to follow, the more employee behavior there is to monitor, the more customer and employee transactions there are generating data — the more you need automated software to monitor compliance.”

Network-planning roles are intensive in data collection and processing, as are scheduling, pricing and revenue-management functions. In fact, pricing and inventory analyst roles are so heavy in data collection and processing, one might wonder if they could be completely automated using machine learning.

If the decision to price or modify inventory availability is based solely on historical data and what the competition is doing today (more data) and the decision is either to match or slightly undercut the competition (simple rules-based decision), then couldn’t these roles be 100 percent automated?

Alex Cruz, CEO of British Airways, thinks so. He has told members of the airline’s revenue-management department that they will be out of a job in the future as the volume of data available and the manner in which it is leveraged to dynamically price on a per customer basis is too big to be processed by humans.

In the back office of network operations, controllers make decisions on which aircraft to assign and reassign to flights, which flights to delay and which to cancel to keep the overall network running on time and successfully delivering as many passengers to their destinations as possible.

Is this not just one large pool of data that machine learning could be applied to that would almost completely automate these roles? Are load controllers really needed to calculate and electronically distribute a load sheet and, equally, are flight dispatchers required to calculate and distribute flight plans?

Airlines often justify the return on investment for new technology tools to assist these roles, based on the number of load sheets or flight plans they can generate per shift. The reason airlines continue to invest in these tools is because each new generation of tools is automating more and more of the process. At some point, the processes could be fully automated.

WITH AUTOMATION COMES MANY BENEFITS, SUCH AS INCREASING PRODUCTIVITY AND FREEING EMPLOYEES FROM SPENDING TIME ON MONOTONOUS TASKS SO THEY CAN FOCUS ON MORE CRITICAL RESPONSIBILITIES. HOWEVER, THERE ARE ALSO DISADVANTAGES TO AUTOMATION. FOR EXAMPLE, THE MORE ADVANCED THE AUTOMATION BECOMES, THE LESS ASTUTE THE HUMAN OPERATOR ALSO BECOMES, WHO MAY FALL OUT OF PRACTICE SHOULD A TECHNICAL FAILURE OCCUR. THEREFORE, IT IS ESSENTIAL THAT TRAINING IS MAINTAINED TO ENSURE HUMAN INTERVENTION MAY BE PRACTICED IN TIME OF A MALFUNCTION.

However, it is not enough to consider only the technical feasibility of automation but other factors such as the supply and demand of labor must be considered. If labor is low-cost and plentiful, there may be no justifiable business case for automation — even with output quality, speed and volume increases taken into consideration. Some regulatory frameworks will not accept removal of humans from certain processes, and although regulations can change, it is often a slow process.

Then, of course, there is the human factor. Would airline passengers, for example, accept a mid flight meal service from a robot moving down the aisle as opposed to a friendly flight attendant who can offer them a more personalized, human experience?

Maybe they would and this would be yet another way that low-cost and full-service carriers could differentiate themselves from each other — one that offers a human touch and one that does not.

The Paradox Of Automation

The paradox of automation is that the more efficient automation becomes, the more critical the human operator also becomes.

“It applies in a wide variety of contexts, from the operators of nuclear power stations to the crew of cruise ships, from the simple fact that we can no longer remember phone numbers because we have them all stored in our mobile phones, to the way we now struggle with mental arithmetic because we are surrounded by electronic calculators,” said London-based economist and journalist Tim Harford. “The better the automatic systems, the more out-of-practice human operators will be, and the more extreme the situations they will have to face.”

Harford illustrates the three factors of the paradox of automation.

“First, automatic systems accommodate incompetence by being easy to operate and by automatically correcting mistakes,” he said. “Second, even if operators are expert, automatic systems erode their skills by removing the need for practice. Third, automatic systems tend to fail either in unusual situations or in ways that produce unusual situations, requiring a particularly skillful response.”

DATA COLLECTION AND PROCESSING REPRESENTS ONE-THIRD OF ALL ACTIVITIES PERFORMED ACROSS JOB FUNCTIONS THROUGHOUT THE UNITED STATES. BOTH ACTIVITIES, AS REPORTED BY MCKINSEY & COMPANY, HAVE A TECHNICAL POTENTIAL FOR AUTOMATION SURPASSING 60 PERCENT. THE IMPACT ON AIRLINES IS EXPECTED TO BE PROFOUND, CONSIDERING THE NUMBER OF ROLES ACROSS AN AIRLINE’S BACK OFFICE THAT HAVE A HIGH PERCENTAGE OF ACTIVITIES RELATED TO DATA COLLECTION AND PROCESSING.

If this paradox is applied to the automation of pricing or inventory-analyst roles, what might happen if the automated systems break down? How would analysts, who have only ever supervised a fully automated system, know what they should do? In a network-operations center, would those who supervise the automated systems performing existing operations-controller, load-controller or flight-dispatch activities know what to do if these systems broke down but all they ever knew was how to supervise the automated systems? Many airlines have been addressing this issue for years in a cockpit scenario. The safest airlines in the world ensure their pilots maintain the highest levels of proficiency by requiring them to y a certain percentage of flight time, takeoffs and landings manually. This ensures, if the automation fails, the human pilots are still at the top of their game if manual intervention is necessary.

Organizational Creativity

While research suggests that less than 5 percent of roles will be completely lost to automation (based on current technology), in the next decade or two, there is no doubt the impact on automation will be far and wide. Nearly all role types will likely be impacted in some way, to varying degrees.

Automation will require many employees to gain new skills related to the supervision and tuning of automated systems — including how to recover them from the occasional failure.

There will be many positives to gain from automation, too, where its costs are justified by its benefits. Airline employees will be able to focus on higher value work, such as serving customers, which contributes higher returns to the airline and its customers. This work is likely to be more interesting than the repetitive work of old that has often made employees feel like they are robots. It is time for airlines to leave that work to the robots and unleash a new wave of organizational creativity.