Decision making for travelers: Are machines up to it?

Decision making for travelers: Are machines up to it?

An interesting result from a recent McKinsey Global Institute study suggests that a typical business leader spends 25% of his/her time on gathering/analyzing inputs for decision making such as reports and data - a task that can be automated using currently available technology. Yet, a large part of the value addition by the machine is in processing large quantities of information and distilling it down to what's important, rather than taking an actual decision and being accountable for it.

To illustrate, a typical traveler must find answers to these questions (among others) to make his/her trip meaningful:

  • Where do I need to go? (Destination)What is the best use of my time there? (Objective)
  • How do I achieve that objective? (Itinerary)
  • When should I be at my destination, to enable that? (Schedule)
  • Why is a certain itinerary my best choice for this? (Confirmation)

Based on the decision-making capacity of the computer, and what knowledge the traveler would have without the help of the computer, it is vital to decide who must answer which of those questions. Look at these three scenarios: 

Scenario 1

  • Customer: I want to attend the ZYX innovation conference. Okay, I see that it is in London from the 1st to the 4th of April. I'll fly British Airways on March 31, because I like to be there early rather than landing on the day of the meeting. Good, there is a flight at 8 pm. I'll return by the 6 pm Lufthansa flight on the last day of the event. I'll need a room at the ABC Residency hotel, because it is the closest. Also airport transfers and cabs to and from the venue.
  • Computer: Your booking is confirmed

Scenario 2

  • Customer: I want to attend the ZYX innovation conference
  • Computer: Certainly. The event is scheduled for April-1-4, 2017 in London. Based on your calendar schedule and preferences, I would recommend that you take a British airways flight on March 31 at 8 pm and return by the Lufthansa flight on April 4 at 6 pm. Please confirm, so I can book your tickets, hotel and cabs.

Scenario 3

  • Customer: I want to attend the ZYX innovation conference
  • Computer: Negative. I have analyzed the guest list for the ZYX innovation conference in London and based on your sales record, I estimate that you will have only a 20% chances of improving your pipeline by attending the event. Instead, I would recommend that you attend the domain conference happening at the same time in Beijing. Based on the guest list and from analyzing your correspondence with Example Corporation, you have a 70% chance of signing a deal with them. I am booking you on the China Airways flight on April 2 at 11 AM. I shall update your calendar accordingly.
  • Customer: No, no…override! Override! I'll make the decisions here!

To look at a very simple model, a decision is the result of processing available information in a meaningful context. A 'bad' decision in this scenario would be one made by the computer which doesn't match the customer's ultimate choice. This stems from either incomplete/inaccurate information or from using the wrong tools to process the right information. It is possible that the customer may often have information or preferences which wasn't properly communicated to the computer because it is too intangible in form to be fully comprehended.

 

The consensus today is that Industrial Revolution 4.0 - the crux of which is universal connectivity or the Internet of Things (IoT) – is upon us. Sensors today can pick up signals which humans are fundamentally incapable of, while devices can talk to each other and exchange information like never before. In other words, machines have a clear edge over humans when it comes to gathering information from a well-defined source. On the other hand, when the parameters are hazy or there is a lack of tangibility when it comes to the context, humans have a definite advantage over machines. A machine can be (easily) programmed to make a choice from the options available. It is very easy to specify to the machine that it must choose the cheapest flight, or to add a handful of tangible conditions – such as shortest flying time, least number of stops, etc. – to the mix and prompt the machine to make an appropriate decision.

For instance, the customer may consciously choose not to fly with a particular airline or not to stay at a particular hotel for a few months because he/she had a prior bad experience with them. Or he/she may choose to take a flight one day later than the computer's choice because he/she wants to attend another event in the city or squeeze in an extra meeting before returning. These are neither perfectly tangible, nor are they permanent conditions which can be coded into an algorithm for the long term. On the other hand, a notable discount in prices, or a recent change in a hotel/airline's customer service principles might cause him/her to drop the other preferences and go with a previously least-preferred airline/hotel.

Therefore, it is perfectly acceptable (at present) that the computer is unable to make the 'right' decision with what was available to it, as long as it records the customer's preference and adapts accordingly for future decision making exercises. But that also comes with a caveat – that the reason is tangible enough to enable proper adaptation. Once the booking is finalized, imagine that a simple popup comes up:

Your choice did not match my recommendation. What happened?

Until and unless the customer can clearly articulate the answer to the above question, computers will have to wait until they take over all decision making functions from humans. Ironically, that brings us to the biggest decision of all: what exactly do we want machines to do? And that is the one decision we have got to make ourselves!

Rate this blog entry:
0
Distributed, scalable airports: Shorter drives, fe...
IBS partners with Traveltech Lab to boost innovati...
 

Comments 1

Guest - Lekshmi Raj S on Thursday, 16 March 2017 12:40
Lekshmi Raj S

Good one Aravind!
I believe the very same question: "what exactly do we want the machines to do?" ripples in many minds.
Thank you for sharing!

Good one Aravind! I believe the very same question: "what exactly do we want the machines to do?" ripples in many minds. Thank you for sharing!
Guest
Friday, 28 July 2017

Captcha Image