As airlines emerge from the COVID-19 pandemic, what trends will outlast the recovery?
The pandemic was the single most challenging period for the airline industry, but equally it has proven to be the most opportunistic time in terms of adopting new, more modern and innovative technologies.
In particular, the pandemic highlighted the need for airlines to be prepared for periodic and unpredictable ‘demand reduction’ events and ‘demand enhancing’ events. Key to this is the need for a high degree of flexibility and agility to react in real-time. Examples include real-time dynamic pricing and offer management that reacts immediately to ‘demand impacting’ events.
Demand profiles from the two primary traveller market segments – leisure and business – have changed dramatically with business travel experiencing structural change, while conversely there is huge pent-up demand for leisure travel. While this bodes well for airlines in terms of traffic volume, leisure travellers are more price conscious and seek lower price points than more profitable business travellers. Airlines will have to maximise pricing to ensure the highest possible revenue is being generated in each ‘shopping engagement’ without impacting demand.
The pandemic also shone a light on issues associated with legacy systems underpinning airlines’ technology stack, as these systems were slow to react to rapidly changing environments.
The opportunity for significant change in airlines’ underlying technology, particularly leveraging AI, automation and SaaS is huge. And the opportunity is now.
You are leading AI initiatives at Datalex to transform airline retail, particularly in the area of pricing. What do you see as some of the key ‘AI’ opportunities?
We see several areas where AI could have immediate transformational impact in the travel industry.
One key area rife for disruption, leveraging AI, is pricing. The pandemic highlighted the constraints of the traditional pricing approach which is restrictive & static, limited to specific customer segments and demand scenarios, limited price points (via RBDs), that cannot adjust pricing in real-time.
With the traditional revenue management set up, airlines suffer from a trade off between ‘price accuracy’ and ‘speed to market’. For an airline to update its price predictions frequently, it must analyse less inputs and vice versa, if it wants to analyse more, it has to update predications less frequently.
AI is fundamental to overcoming this trade off by cultivating pricing strategies that adapt in real-time. RBD-less pricing, powered by AI is the future. This will allow airlines to understand and react to fluctuating market conditions, competitive landscape and other complex data which requires sophisticated data analytics and machine learning capabilities.
Customer service and end-to-end customer engagement throughout the travel lifecycle are other key areas. AI can be used across the board to enable airlines to be more reactive and engaged with travellers without the cost of implementing in-person engagement. Sophisticated, AI-enabled chatbots can be implemented to quickly resolve passenger queries, improving operational efficiency by reducing call center load and increasing customer satisfaction with real-time engagement.
Ultimately, airlines have a great opportunity to harness AI to transform how they do their business, to connect better, and with more customers and to drive new revenues
It’s an exciting time to work in this industry where real change, and at a pace never experienced before, is happening every day.
What do you think would be the ‘quick wins’ for airlines in their AI strategy? And what should be on their medium and longer-term ‘AI’ roadmap?
As described above, in the short terms airlines should focus on pricing and customer service.
The total cost of air travel has the most significant impact on a person’s decision to purchase an airline ticket. The likelihood of delays and cancellations, and loyalty status also play a significant part in the decision-making process. Airline customer service is crucial to support customers before, during, and after a flight. Improving customer service leads to happier customers, better travel experiences, and higher revenues. These are the areas that airlines should make a strategic priority when it comes to “quick wins” using AI.
AI-powered customer service chatbots is a quick win that delivers high impact. Chatbots provide significant customer support savings. Second, bots can also sell.
Another key opportunity that remains untapped by airlines is AI-powered dynamic pricing which has the ability to drive higher revenues for the airline and satisfy the pricing needs of the customer through smarter, reactive pricing based on actual real-time demand and customer price sensitivity.
Ultimately, capturing more customers and more opportunities in real time across the revenue cycle.
Personalised offers leveraging AI should be a medium-term focus for airlines as this will enhance the overall customer experience. Airlines can use AI to learn about the behavior of passengers and create offers that will convert better. We refer to this as product determination.
A longer-term focus should be on AI driven dynamic offers and eventually a move to ‘one order’, and those initiatives are coming down the track sooner than we think.
Furthermore, AI can hugely improve operations below the wing.
How else can AI powered dynamic pricing benefit airlines outside of additional revenue?
This is a great question, because this topic rarely comes up, and it’s a topic I’m really passionate about. Besides the obvious revenue uplift, AI-powered dynamic pricing also can generate significant operational efficiencies for airlines.
Revenue management’s goal is maximise their revenue across the entire network, with the “entire network” being the key phrase here. Once pricing strategies are defined and fare structures filed, revenue management will then use a combination of their RM tools to manually monitor market conditions.
There are only so many markets revenue management has the capacity to monitor and there is only so much time a human and the processes allow to react. As mentioned earlier airlines suffer from a tradeoff between ‘price accuracy’ and ‘speed to market’. AI powered Dynamic pricing can be used to offset these workloads, react more accurately in real-time and direct Revenue Management’s attention to strategically focus on the markets with the most opportunities, and other value driving tasks.
What other ways can AI be applied within an airline? What would you suggest are the top 5 ‘AI opportunities’ for airlines?
The adoption of airport robotics, blockchain technologies for data sharing, and VR technologies will be widely used in the future. Blockchain for data sharing was explored by Datalex in a recent hackathon.
AI is already being used to estimate the average lifespan of the parts on an aircraft. AI also has the potential to be fully integrated into the flight scheduling process to reduce offered flights during a significant demand restriction event.
During the pandemic and because of the route/gate rules airlines were frequently flying very low occupancy planes or empty planes on routes to meet requirements – these were referred to as ‘ghost flights.’ AI could be used to quickly and efficiently manage flight offerings based on demand. When a demand restriction event occurs, zero booking flights could be quickly cancelled and removed from the airline’s schedule. Low booking flights could be cancelled, and passengers automatically rebooked to other itineraries that meet their needs to consolidate more travelers on fewer flights.
What are the top 5 AI opportunities? I would say:
- Revenue Management
- Customer Service
- Network Planning
- Crew management
- Air Safety and Airplane Maintenance
What are some of the biggest challenges in applying AI in the airline industry? Are there any tips you are willing to share?
Applying AI to the aviation industry is an inevitable transformational change that is just beginning to take off. The biggest challenge faced is the lack of knowledge around AI and the true value it can add to an airline, so that educational piece is really important.
Another big challenge with AI is transparency, specifically the fact that humans should be able to form coherent explanations of the reasoning process in AI. That means providing airlines with clarity of what decisions were made by the AI technologies and why they were made. This is hugely important for airlines to understand that AI can be tailored to their business models. Transparency is important not only for trust, but also for debugging, testing and certification.
Another big challenge in applying AI, particularly in the airline industry, is the underfitting and failure to generalise Underfitting is defined as when the machine learning algorithm fails to learn enough from the training data during the training process. Failing to generalise occurs when the model successfully fits the training data, but has a high error on the validation/testing data. This is a challenge that we have successfully addressed in Datalex.
What tips am I willing to share? It requires specialist AI talent and technology partners focused on AI. When you are working with AI, you have to enjoy the process just as much as the achievement.
Datalex’s purpose is to transform airline retail.
To learn more about key digital retail trends and AI-powered dynamic pricing opportunities, download the Datalex Research Report: The Digital Airline and Customer 2022.