8th March, 2022
From being operations-centric to customer-centric, how to cover this distance?
Airlines, as an industry, are often posed this question.
Considering that almost all the carriers have gone through a prolonged precarious phase, lasting almost two years, what would it take to chart a promising future for airlines that is all about what passengers want, whether they fly or not. How to let them be in control when they use a digital touchpoint?
The key to being customer-centric and accordingly crafting a desired revenue model lies in making the most of data that is available. But how to get closer to travellers by knowing them better and building on a proposition that is not only about travel and mobility, say going from A to B in general, but also encompasses a traveller’s lifestyle.
As a visionary and an advocate of the path that airlines need to focus on, Ricardo Pilon refers to facets that revolve around travellers. They are as follows:
Three pillars to being customer-centric: Pilon asserted that there are three pillars that drive how you become customer centric, and be relevant in each changing context.
- We must become more intimate with the customer about their purpose in seeking experiences
- We have to allow them to show us how they think, and how they make choices and trade-offs
- Once we have that, we have to solve problems and make it easy for them to use it.
“This is more important now because we must remove duplication of efforts. Loyalty has allowed us to get to know people better. As a higher proportion of individuals are frequent travellers, we can start to do more with the data. We have sufficient mass and technology to elevate the business model now,” he said. “But the next step is actually driven by the customer, underpinned by deep learning and Enterprise-level AI. With permission. Loyalty should drive this, and ultimately it will be better equipped to embrace customer-level revenue management in a new function.”
Understanding the intent of the traveller: According to Pilon, a customer can simply tell us in human language what their intent is. We can use Natural Language Processing in AI to take that further. If they cannot articulate it in words, then we use deep learning to derive it. “But, as said, customers will be part of it. I do not believe we can solve all problems by deriving insights. It also does not help people adopting it. Further, complex human data is like the principle of supervised learning in AI: It needs help in where to look and how. So, we need to co-create, but use automation to solve human thinking with more computing power in real time. But it must follow the neural network we as people use,” he said. “What I describe is really how branching, loop, convolution algorithms are written. By humans. They simply help us achieve our results faster and only use the artificial intelligence to execute what it is we want it to achieve. If those goals are defined in customer objectives, we bring benefits to people. They will adopt it, and they will pay for it.”
Serving passengers
So from a passenger’s perspective, airlines must look at how consumers plan a certain experience.
For this, airlines must be less dependent on derived demand (air transport). “The pandemic made this visibly painful, but we also said this after 9/11,” said Pilon.
If we look at the current booking path, which tends to originate from a flight departure button, it tends to limit one’s horizon, doesn’t let one express what they are looking for and often ends up making the booking path painful.
“…it is not even how most experiences are planned. Again, the pandemic has exacerbated that, too. Now, we even worry about restrictions overseas,” said Pilon. “I am advocating that we connect better to individuals and allow them to plan the experiences and activities they seek and fill in the gaps after. For example, once I know you, and you choose to combine cooking classes with local chef Khmisa and visit the Ouzoud Falls outside Marrakesh in Morocco, I can make everything else easier for you, and even help you choose between options on how to get there, move around, all based on your own rules and lifestyle choices. Better yet, I can help you and optimize value for your budget overall. Based on your rules. The same happens with attending a conference overseas. I know when it starts. Why do I have to figure out when I have to leave my hometown and how to get there in time? The more I travel, the more time I waste. That does not work in a world of Internet of Everything (IoE).”
Being data-centric means centricity in goals
As Pilon pointed out, today, the term data-centric is more used in a departmental context. For instance, counting predictive analytics to better forecast demand and seat selection fees.
“But airlines are not truly data-centric until they are goal-centric. Even airline balance scorecards have dimensions that cannot be aligned or integrated into a clear KPI. What does that say about organization design?” highlighted Pilon. He continued: “So, number one, we need to be able to define our goals better around our customers as individuals. Not passengers. We are not clear about this. Are we transactional, opportunistic, or are we building relationships? Is what we do air transport from A to B? Or are we solving your problems every time you need us? Being data-centric also means centricity in goals. And unifying your data sets around a goal. It also means making choices. Flight based? Customer based? If airlines truly want to transform their service offering, we may have to sacrifice short-term pains to cure the fundamental flaw in our service industry. All departments pull the goals and customers in various directions. That is why that becomes especially painful in an era of IoE. That is, the customer is now acutely aware of it!”
Pilon explained:
- The unified goal must be defined in customer terms while delivering toward business objectives. That could be the highest ‘completion factor’ of meeting customer goals (solving their experience planning problems based on their perceived value/KPI) while getting the highest share of wallet from each customer each year (not each flight). That could be the business KPI. But goals are choices.
- Once you have defined those goals, you can write the layers of logic in how you achieve that. That is the execution of all your processes in a value chain across all departments the way the customer experiences it. In layman’s terms, it walks the customer from the outside into your organization and has a consistent experience through their PDA at any point in time, anywhere. It does not matter there is loyalty, marketing, or RM behind it today. It is a customer process. They should not be exposed to it.
- You can automate how these processes are executed with Enterprise AI. But they all have cost, value and actual price dimensions today. Because we still have these departments. So, you stack multiple micro-optimization layers onto each other to solve the overall puzzle. For example, you do not want to start charging more for checked luggage to a customer you see will be top tier status next month, even though for a transactional customer, you could do what you want. But that depends on your customer goals, too. Every process is stress-tested against the goal. There is no other way. Layering optimizers on top of existing processes, like seat pricing, allows you to respect their lower goals but make it consistent in a bigger picture. That is how we can offer discounts on seats based on how much else will be spent. So in sum, all these relationships must be laid out, before you write the end model with all algorithms to execute it in real-time.
From data to a retail roadmap
Ideally one is looking at a data platform that can ingest data from a variety of sources and come up with suggestions with the right mix of content and products.
Pilon spoke about a concept called “retail marketplace of experiences”, toward which he is developing a solution.
“In its simplest description, it’s a digital assistant that is designed by the customer that is using it,” he said. The way it solves problems reflects the thought process and choices they make. It allows the customer to bring in its influencers, decide what content to trust, store playlists and preferences but allows the tool to do the heavy lifting in solving the puzzle. “It is a trusted assistant because it is neutral, too. Third-party content the customer chooses can be brought in through data streaming but is commercially bundled in a convenient fashion. Any changes automatically retrigger other recommendations and updated prices. It displays alternatives that optimize the spend as per the budget. In the background, the activities are serviced by the relevant stakeholders. All other duplications can be removed. This is about making your life easier,” explained Pilon.
As for the retail roadmap, it initially builds on today’s ancillary strategy. “It is by adding related products and services, such as excursions, golf tours, anything lifestyle related. But ultimately, we will hit a natural and psychological barrier. Customers do not want more content pushed to them in the same way going forward. And most retailing programs are centralized in that they show ‘own’ or merchant content. Not everybody trusts the content and pricing, or the way it is presented. Even “superapps” are not necessarily built around the customer today. They just integrate more in a single App. That is not the same,” he said.
The gap, according to Pilon, is bridged by layers of deep learning and departmental optimizers into a commercial layer where the use case is the customer’s budget, their goals, and how the digital assistant solves that.
Pilon mentioned that it is not a blackbox of automation.
“This is the best combination of human intelligence, and treating customers as human beings we want to have a relationship with, but using computing power and technology to solve it in real time. That is bringing benefits to people. That is how you are allowed to charge for it,” summed up Pilon.
By Ritesh Gupta
Ai Editorial
