The online retail industry continues to grow in India. According to analyst firm Forrester, the value of online sales in the country is predicted to reach $48 Billion by 2020, even after taking the effects of demonetization into account. It’s clear that the sheer size of the market, of which a substantial part still remains untapped, and aggressive pricing tactics are going to push sales numbers upwards. But from a technology perspective, it’s differentiation in service, user experience and automation of back-end processes that will play an important role in defining market leaders, and this is where the concepts of artificial intelligence or deep learning come into picture.
Let’s take the case of product images, given the magnitude and complexity of the problem. Today, online retailers and marketplaces with medium to large catalogues are required to manage a large number (millions) of products. A herculean task often riddled with manual steps prone to human error. One of the most crucial steps, and one that directly impacts search results and in-turn sales, is that of tagging - the backend process of adding descriptive text to images so products can be indexed and searched. Currently, this is a manual process for most online retailers, which means that it’s expensive, time consuming and error prone. More importantly, inaccurate tagging results in frustrated users and lost revenues. But, fortunately, there’s a solution to this problem that retailers around the world have started to adopt. This solution I speak of is artificial intelligence.
An AI powered approach, that takes advantage of deep learning and computer vision, helps retailers dramatically reduce the amount of time it takes to tag products while improving accuracy. Instead of manually tagging images one by one, retailers can upload all their product images to an AI engine and have them tagged by color, pattern, style and other parameters all at once. Similarly, Artificial Intelligence can also being applied when it comes to making intelligent product recommendations. In this use-case, depending on what shoppers are looking for at a given time, appropriate suggestions for visually similar products can be made, without having to link products manually in the back-end.
In addition to improving backend processes, retailers across the globe, including a few in India, have also started to use artificial intelligence to transform the way in which users look for products. Retailers can now enable online shoppers to find products by uploading an image, instead of searching for it by text. This means shoppers can now find products they saw in a magazine or on a social network, without having to guess what the right search keywords for that product might be. This is especially effective when it comes to categories like fashion, where searching for styles and colors can be challenging via conventional text based search. For retailers, it’s an opportunity which helps them tackle the problem of discovery while building new experiences.
We know that online retail will remain a hyper competitive space, one where building unique experiences or customer loyalty is always an uphill task. But by using the appropriate tools such as artificial intelligence to solve challenges related to images, retailers can remain ahead of the curve when it comes to driving efficiency and revenue.
The article has been penned down by Anmol Arora, Head, Visual Commerce Platform, ViSenze