Human at the center

Hyper-personalization is a an important trend of the market used for different business goals  (like friction-less experience featured in this post )

However, in industries like travel, even if an important part of the data processing can be managed machine to machine, the goal of the personalization is first of all to increase the quality of the human experience so the systems will have to foster machine to human and human to human interaction.

But how to define a personalized experience with a passenger that will most of the time flight for the first time with the crew and event with the airline.

Like Andy in “The devil Wears Prada” during business receptions, the system needs to give the information can be gathered to the crew so that the best interaction, based on cultural intelligence, can be created toward the passenger.

What we have learnt from the past is that direct inappropriate use of the information (like the public display of the traveler looking for incognito like celebrity, or publication of data consolidation like pregnancy detection) will lead to discomfort and produce the opposite effect than the one targeted.

As emphasized by many books (like the good “Effective Communication Skills” from professor Dalton Kehoe),  our communication is a complex model that rely on the understanding of the value of the recipient.

Also, the culture intelligence, based on awareness and understanding of the cultural clusters in the world, needs to leverage the interaction towards different variations (direct/indirect relation, distance …)

More information on CQ with the book of David Livermore, Leading with cultural intelligence

Personalization have been more powerful with the help of digital services and big data, as the ability to collect data on the user increase the probability to have a beneficial impact on the traveler.

Those interactions, managed by the crew in their daily interaction, will happen in 3 dimensions : 

•  Physical
• Communication
• Culture

If we look, in at a futuristic approach, what could be simplified with the insertion of technologies ?

For physical, robots can operate autonomously in wide space, but close space like aircraft increase the complexity of the physical interaction (like grabbing a glass on the floor).

Some recent examples like the ALIAS co-pilot demonstrate that human interaction can be partially replaced in context but we are, for the moment, not ready from an application in the cabin.

ALIAS flies and lands simulated Boeing 737

https://www.dailymail.co.uk/video/sciencetech/video-1466173/Robotic-pilot-ALIAS-flies-lands-simulated-Boeing-737.html

For communication, the automation implies to detect the current spoken language but also the accent (that can imply another native language)

From there, based on the idea of a butler and the model of habit (a set of stimuli trigger an action), several concepts have been developed to automatize interaction : internet bot for browsing, chat bot for interaction, social bot for networking …

With the progress of voice recognition and voice generation (text to speech), the behavior can slowly interact directly with human like the example of the pizza order by google

Google I/O 2018: A Google Assistant that will even make calls for you

https://www.youtube.com/watch?v=d40jgFZ5hXk

The value of those services is created by transforming and connect the data information(or the data source), see the talk about the 3 models of data for more example

For culture and mindset, we are making progress in detection of mood (facial or voice modulation), understanding of usage (with CQ), automatic control of content compliance (like nudity and violence with machine learning). But, this aspect will be more slow as it needs a structure meta-model

One of the core difficulty of the automation is the model; even if we could collect all historical data and information on the situation (like identity and travel plan), the behavior of most of the algorithms relies on the ability to create a model of the situation and to define an appropriate reaction to it.

The panel of solution is wide :

The situation synthesis can be created with various level of analysis and complexity : based on trigger on a specific threshold (like flat pressure), a divergence from average (like predictive maintenance), a keyword correlation (like medical diagnostic), pattern matching (like road driving or human detection) or more complex algorithms (like morpho-mathematic for OCR).

From there, the action model will use deterministic algorithms (like decision tree) or learning algorithms (like neuronal networks) to evaluate the appropriate decision.

Finally, the execution based on the information of the predictive model can be apply with direct visibility to the customer (like a personalized landing page) or indirect (like boarding order of passenger, optimization of the loading of the food tray, or priority to address customer needs)

As we focus on the seamless travel experience for the passenger, the system needs also to assist the crew in their operation.

A similar principle than seamless travel applies then to crew personalization which is looking for seamless operation while navigating from one aircraft to another or across the airport.

As demonstrated with several studies (like the amazing book of Daniel Kahneman Thinking Fast and Slow) the reconfiguration of the processing part of our brain (system 2) for a new task is a difficult and energy consuming activity.

By combining information in a way that allow the crew to focus on similar actions and to anticipate next phases, the system can provide an important release and even simplify the procedures by “just in time” or contextual display of information.

This approach have also been used in the re-design of the cockpit in the 2020 project of Thales

See video on this link 

However, as the connectivity and mobility equipment (like tablet or mobile) make their place in the eco-system, the crew will be faced to the same dilemma than doctors a few years ago.

On one side, the electronic device give them information on the patient, but on the other side, it also create a distance in the relation :  as screen on the desk can create a physical visual barrier or force them turn his back to the patient.

A successful combination of layout arrangement (like tablet on the tray), use of wearable devices (like smart watches) or integrate insertion (like pico-projectors) can restore the direct relation.

But this environment adaptation implicitely create the need for a dedicated pool of devices asserved to the crew needs.

To realize it, the system has to evolve to embeed a full EMM component (Enterprise Mobility Management) that will allowto update, configure and manage security of those dedicated devices (for both crew and passenger identity).

This functional IOT grid of components will be a closed extension of the rest of the aircraft cabin system but with, most of the time, an easier ability to integrate up to date technologies for behavior and exchanges

The grid will have 2 missions :

• connect to the passenger IOT grid to achieve the hyper-personalization
• connect the crew members together to facilitate the global operation

One of the particularity of the transport industry is that, unlike the public domain, both passengers and crew members are using the aircraft as a temporary environment during a flight.

The high crew mobility is also a reason why the system needs to  facilitate the internal team discovery.

Many airlines have operations organized with a large number of employees (combining member mobility, availability and other constraints like shift regulations) which means that a crew team is often assembled for a short series of flights only

To quickly acquire members specialty (like language knowledge or security expert) can be a differentiation in the passenger relation for both normal (like travel guidance) or urgent procedures (like medical assistance)

The time that the system can save on repetitive actions increase the time available to established personalized and professional connections with the passenger.

A more efficient server to a frequent flyer helping him to use the travel time to be prepared for his next challenge, or to help a single parent in challenging situation with a young baby, are examples of support that create a difference for more than one flight or more than one passenger.