Hi, How May I help you?
Last weekend, I watched a movie which was about how consumers in one of the states of India were not getting proper electricity services in their villages and cities. And when they tried to connect with the authorities, their complaints were not answered as well. During these moments, though my eyes were focused on the TV screen, but apart from processing the movie frames, my brain was also busy in retrieving Datacated weekly’s NLP topic from its amygdala. I was thinking how NLP can be used to answer their calls and emails, as there are not always enough representatives in times of crisis like this.
After spending some time on Google, I came to know about the term associated with this task, i.e. ‘Customer Service Routing’. It is used for redirecting calls or emails, which a company receives on its end from their customers. Now, if we could use NLP in this, then I believe a lot of time could be saved on both the sides.
If you think in terms of numbers, the most populous state in India have a population of more than 204 million (FYI population of US is 325 million). Out of these 204 million, even if 1 in 1000 is registering a complaint daily, then the number of complaints for a day stands at a 2,04,000. That’s just for one day, for a month it’ll be 61,20,000 and in a year it’s almost 74 million.
I understand it’ll not be that easy to implement, as we can’t expect end-users to use only English as their medium of communication, for that, NLP models must be trained with local languages specific to those states only. But if we could do this, then handling of those humongous numbers would certainly become a lot easier.
By: Amit Bishnoi