NLP in Retail Industry

If you back probably to 10 years ago, most of the applications of NLP in retail websites were limited to indexing store products and search user query on top of it using algorithms such as TF-IDF. The recent evolution in NLP also helped retail industry mostly in two forms: Search & Recommendation

Search

Have you been in a situation that you were looking for a type of product but didn't know exactly its name? or maybe just wanted to ask someone to help you? with the new generation of chatbots and semantic search this problem is much simpler now. Behind the scene your query is converted into a format that is understandable by computer and Natural Language Understanding - an area in NLP. NLU extracts information from your query and according to a machine learning model either retrieves a list of products or just simply direct you to right channels to get response.

Recommendation

I bought a multivitamin recently from an online retail website and all of the sudden I received a list of other multivitamin as suggestion. It doesn't stop here, you might have seen when you watched a movie in Netflix, it suggests you a list of other movies to watch that most likely are in the same genres. This level of personalisation although is not necessary a new thing but when it combines with all sort of metadata like product information, purchase history and perhaps even tens other factors creates a new shift for retail stores to be on top rank and keeps their purchase rate high.

What is next?

How about if we combine search and recommendation together? You search for a product that doesn't exist in the store, can conversation AI be intelligent enough to not simply reply back "We are sorry, item is not available" and perhaps replies back with something like "It seems we don't have it but how about this product ..?" This means to be engaged with customers and let them to decide. Search and Recommendation modules, these two old friend and reunions and build an intelligent system that gives customers an outstanding customer services in all dimensions.