Over the years, the CPG sector in India has turned to big data analytics to unlock insights from text data by scanning the reviews and determining the customer sentiment. Bengaluru-based Turing Analytics founded by Divyesh Patel and Aditya Patadia in 2015 develops Artificial Intelligence systems using deep learning to help businesses make data driven decisions. To begin with, the young startup has three marquee names in its corner to boost its credibility – Shopclues.com, global powerhouse Kimberly Clark and Tata Sons Ltd.
Driving sales and conversions with Customer Review Analytics
In an age where understanding customer and bettering CX has become a science, the founders created their first solution “Customer Review Analytics System” for e-commerce sector that worked amazingly well for both English and Hinglish reviews. Infact, ShopClues became the startup’s first client to successfully integrate Customer Review Analytics to derive more insights from text data. With a strong market for analytics, ecommerce and retail giants are quickly turning to machine learning systems, NLP and data science to detect emotions, entities and understand the real context from text and social data. Today, in a market saturated with customer analytics solutions, companies, big and small alike are turning to startups or established vendors to race ahead.
Pegged behind-the-scene analytics, the underlying objective is to boost online conversions and sales. That’s where Turing Analytics steps in with each of the products having a specialized deep neural network at the core that provides it with intelligence. “In case of Reviews analytics, it analyses language to derive context and in the case of Visual Search and Image tagging, it understands what is there in the image. In case of Fashion Trend recognition system there is a neural net which finds patterns, colors and designs that are occurring more frequently,” said Patel, adding what differentiates their products from market leading vendors.
Here’s a look at Turing Analytics Solutions
Fashion trend recognition – It is the latest product launched by the startup, the founders have developed Neural network based AI that can identify what is trending among customer base and suggest which products should be pushed.
Similar product recommendation – This solution suggests similar looking items from the catalog to the customer. The startup’s deep nets can identify similar items based on only images.
Visual search – This solution enables users to search items using photos and it can identify the right items being searched without cropping or category selection. It finds the right results from millions of images within milliseconds.
Image recognition – According to Patel, the startup trained customized image recognition models as per client requirements. Most off the shelf solutions don’t identify different ethnic wear with as much accuracy and that’s where we come in, he shared, explaining how their solution differs from what their competitor is offering.
Facebook Messenger Fashion Chatbot – An easy-to-integrate fashion Chatbot for Facebook messenger acts as a medium for customers to interact with the brand and ask what they are looking for and make conversions by offering similar products from the catalog. The chatbot is aimed at pushing targeted and promotional messages based on the user’s past behaviour in messenger.
According to Patel, their Visual Search solution has many takers in the market and the technology has a much broader application. “We chose to focus our efforts on E-commerce sector for now. We have a few undisclosed clients with catalog size ranging in millions for whom we provide visual search,” he said.
Understanding the Turing Analytics USP
With a wide expertise in building deep learning models for any given task, Patel emphasizes the young startup’s strength lies in using all types of data (text, image, video etc.) for training neural networks. Which gives them the added advantage of eliminating limitation — there is no industry or sector limitation for the solution which means what works best for fashion can also work for home furnishing with equal efficiency and for medical images and patent search. “We can accommodate any new client without having to make many changes in our system while providing highest level of accuracy,” he said. Patel cited a use case of the highly appreciated “Fashion Trend Scoring system”, developed using Deep Learning for Group Technology and Innovation Office, Tata Sons Ltd, this software was capable of determining the trendiness score of a T-shirt with advanced deep learning algorithms.
Well, with more and more companies grappling with data deluge and customer centricity driving businesses, artificial intelligence presents a tempting opportunity to rise above the competition or risk falling behind. The seven-member team of Turing Analytics likes to focus both on features and driving better CX, which is evident by its client base. Another trend gaining prominence is visual search and ecommerce giants Flipkart and Ebay have plowed massive investment in this technology and also recently enabled visual search over their global catalogue. The young startup has trained its eyes on this space.
Try deep learning using MATLAB