One of the top priorities in our quest to strengthen our position as a truly B2B4E platform is to keep improving the personalization of our services so our clients and their traveling employees can benefit from a seamless experience.
Over the last decade, huge advances in artificial intelligence-driven techniques like machine learning and deep learning have made it possible to crunch an enormous amount of data, both numeric and non-numeric, and generate a highly-optimized outcome. Put simply, and in the context of travel, this means we now have the tools to assist in the prediction of travelers’ needs at each and every point in their journey with enormous accuracy. This might sound simple to do but is quite challenging to accomplish.
Let’s use an analogy: If a person orders pizza for dinner three nights in a row, it might seem logical to predict that she will order pizza a fourth time as well, right? Not quite.
There are several reasons why this prediction is fallible. First, just because someone ate something a few nights in a row doesn’t mean they will repeat that endlessly. So, the best we can do is to assign a probability that this person will be ordering pizza. Secondly, it’s important to know the context. Did the person order pizza first three nights in a row because there were no other dinner options around? Or did the person have a kids’ sleepover? Or has the person been working late and pizza is the only option at that time of the night? The variables are endless.
It is only after factoring in all possible contextual variables that you can determine with very high accuracy what the person will or will not order. In technical jargon this is called multi-variable testing or multi-variate regression. Machines can do it far more efficiently than humans.
These techniques can also be used for many other facets of your travel journey to make the experience more enjoyable and frictionless. At CWT, for instance, we now have an increased ability to predict well in advance of a trip if the flight is likely to be delayed or cancelled with a high degree of accuracy. Based on this intelligence we can – first of all – give a fair heads-up to the travelers and also recommend alternative options tailored to their needs as well as their companies’ requirements.
Another new innovation at CWT is our personalized chatbot, which can notify travelers when it detects there is a booking with overnight stay and the hotel is missing from the itinerary. Not only will the bot notify the travelers alerting them of the situation, it also makes hotel recommendations based on the travelers preferences and their company’s travel policy. The same logic can be extended to predicting and recommending other on-trip services like ground transfer and things to do in a highly-contextualized environment.
All of these innovations are quite a feat given that that each individual can assume an unlimited number of consumer personas depending on the context, timing, and several other factors. To be able to sift through all of these seemingly endless combinations in real-time and home in on that specific need with a corresponding relevant offer at the right time, is the secret sauce. To be able to do this for every single person, every single time, across dozens or perhaps hundreds of possible offers, is the new frontier.
Today, we are getting progressively close to this ideal paradigm where we are able to move away from just offering something routine to being able to anticipate a very focused and discrete need and to make a highly relevant and curated offer toward that exact need at the exact time. In other words, hyper-personalization.
And although there has been remarkable progress in technology and techniques to personalize content, we still have a long way to go to create what I call “empathetic AI” - technology that can sense things and situations. It doesn’t yet exist and possibly may not exist in the near future.
This is where human-touch comes in. We, humans, are astonishingly adept at “sensing” things. Not just stark differences between good mood and bad mood, but even highly-nuanced things like sarcasm, subtle body language, hesitation, eagerness, and the list goes on. Humans add enormous value in such instances, and this is where we can complement technology.
Blog author: Utpal Kaul, Head of New Product Incubation, CWT.