NASSCOM Community Admin

Anticipating Customer Needs in the Age of AI

Blog Post created by NASSCOM Community Admin on Jul 8, 2016

A panel discussion at the NASSCOM Big Data & Analytics Summit, which featured, Anand Chandrasekaran, Mad Street Den, Ashwin Apte, Hike Messenger, Mayur Datar, Flipkart, and moderated by Srikanth Velamakanni, Fractal Analytics.

The core business issues remain unchanged of course. However, in our relentless effort to understand customer lifecycle through customer behaviour, it has now acquired a great deal of significance. Tools like Big Data & Analytics, AI, can help bring about a very high degree of customization, to render a deeper understanding of the customer. E-commerce, is replete with examples of optimization models that indicate product profitability at a granular level, which was erstwhile not possible.

This has led to a data-driven culture. It guides our understanding of anticipation of customers’ needs, including factors that shape buying behaviour. Nonetheless, over time, it is common enough to observe arrogance set in, and individuals exhibiting a certain know-it-all attitude, which even influences key decision making. Not least being the rigidity that comes in its wake. AI has been known to make a mockery of it - the much vaunted experience, gut feel et al. Data driven deep insights can trump gut feel, to throw up learnings which are diametrically opposite to what “experts” would normally pontificate. Unarguably, Artificial Intelligence presents us with a plethora of opportunities, and yet, we should be wary of falling into the trap of glamorisation. An inevitable pull though, which continues to tug away.

Clearly the emphasis is on customer experience. A simulated one can be very powerful and provide a near-exact experience of having actually visited a sales outlet. Today, the proprietor is aware of these things, and tries hard be aspirational in all aspects of providing a unique buying experience to create stickiness factor. Though much progress has been made, and yet, it is far away from actually replacing the physical buying experience completely. In an attempt to delight the customer at all times, there is a strong possibility that the wrong technology may be used, leading to a tech over-kill. The lure of technology can get the better of us, unless we are careful.       

Today, it is possible to access a multitude of experiences even through a smartphone, and it is important for enterprises to be cognizant of this at all times. Moreover, the solution being offered has to be simple. When enterprises are not able to do this effectively through existing models, then perhaps it is time to consider the use of Artificial Intelligence, far more seriously. In e-commerce for instance, the sellers on the platform are all different companies. There’s a lot of synchronization which is required, and that too in real time. Customers want end-to-end clear access across the entire spectrum. There are other challenges as well, like departments even within organisations have different ways of functioning. It is anything but uniform. The other challenge is getting the right talent.

Smaller organisation may not have very large data sets to be able to leverage full potential of such technologies, and typically the trade-off is between returns and investment. It may be a long drawn affair and if there isn’t a proper plan of funding such initiatives for sustenance, it can all end up as sunk costs. The focus is on customer life time value, but if the funds are not in place, leaders would be compelled to resort to gut feel and common sense.

Finally, we come to the all-important question of ethics. The sheer volume of data that is being shared, including the level of granularity, means that there is a risk of individual privacy being compromised, unless stringent checks and balances are in place. It’s a grey area. Well, if there are no easy answers then that’s what we generally like to say. Enterprises must be mindful of the fact that they are repositories of personal data, which ought not to be misused. In case of a deviation from this, the response time has to be very quick. A proactive approach from the org, is what is required.   

Outcomes