Ever since its inception in 2010, Indix’s AI platform has been enabling industries across the globe to unlock value from product information. They have helped clients improve their product search, automate catalog management, optimise promotions and offers, enrich product pages, and analyse the market and competitors, among others.
Based out of Seattle, Chennai and Hyderabad, Indix recently announced a broader availability of their AI-based prediction services for product categorisation, brand identification, standardised attribute extraction and product clustering. With the help of Indix’s prediction services, companies can now process millions of product records in minutes using AI, without any human interference. This is exactly what Indix has been working towards, since the very beginning — investing in artificial intelligence and machine learning to handle huge amount of information, says Sanjay Parthasarathy, founder and CEO at Indix Corporation.
In February this year, Indix also announced that their AI-based platform has been integrated into Samsung Mall to help customers discover and shop for products across various online shopping sites using Samsung Galaxy On7 Prime’s advanced camera and microphone. During the launch, Parthasarathy had said that AI-enabled, contextual product discovery and shopping, integrated into every smart device has been a long-standing vision at Indix.
As the startup is witnessing an accelerated growth in AI, Analytics India Magazine caught up with Parthasarathy to understand how companies are making use of AI solutions by Indix, the idea behind founding the startup and their future plans. With a team of 55 in Chennai and Hyderabad, and 15 in Seattle, Indix, just like their completely automated information platform which learns continuously, are on a path of continuous growth.
AI Solutions By Indix
Parthasarathy explains that by using Indix prediction services, e-commerce marketplaces can clean and verify seller content, expand product listing attributes and accelerate SKU on-boarding. Adtech companies can standardise advertiser catalogs, enhance ad relevance and power dynamic ads. Affiliates and publishers can automate catalog management and power product search. Market Research and analytics companies can identify and classify products by simply using the information on electronic receipts.
Evolution Of Mobile Shopping Through Indix’s AI Solutions
As shopping is moving towards contextual product discovery which is integrated into all smart devices and software, Indix is contributing to this evolution by developing the technology needed to recognise and understand product information using learning systems. Parthasarathy explains how Indix is powering Samsung Mall which is built into Samsung Galaxy On7 Prime smartphone, through its AI solution:
From Keywords To Multi-Sensory Input: Three years ago, the predominant method for finding products to buy was by entering a keyword or search phrase in a search engine or an e-commerce website. Now, it’s likely that the discovery process is initiated by voice (Siri, Alexa, Cortana, Bixby, Google Assistant), via a camera (using product image recognition), or by applying product image recognition or product phrase recognition to photos and text on social networks, videos, chat, messaging and other content. Any device with a sensor (or the sensors themselves) is a point of discovery for products. Therefore, AI services which help in identifying products using different sensors, and then understand what the consumer wants to do with the product, are critical.
From Recommendations To Perfect Personalisation: Amazon has the ideal recommendation method for introducing products based on one’s shopping and browsing history on their site. Spotify and Netflix have pushed the envelope of personalisation based on knowledge of one’s physical and emotional context. Location-based offers, targeted ads based on internet-wide clickstreams or previous purchases are finding acceptance as well. Consumers, for their part, are expecting to be offered the ‘perfect product’ which matches their current state every time they shop online, or offline. Here too, AI is critical to understanding context and then using product information to present the right product, at the right time, at the right place and at the right price.
How Does The Indix AI Platform Work
Parthasarathy says that Indix platform has three primary components:
- Capturing and processing product data at scale
- Using an AI platform which helps in rapid training, retraining, and deployment of AI algorithms
- Filtering data at a large scale
“Analytics plays a key role in the third process. Once data is cleansed and structured, analytics can be run on this data”, he said.
Idea Behind Indix
Reminiscing old memories, Parthasarathy said that as they got into the space in the early stage of the company, they recognised a few things:
1. “First, products (or things) are one of the essential entities – people, places, things, businesses, documents and devices. Essential in the sense that every device and software needs to be able to understand and do something useful with these entities. Just like they do with location information today.
2. Second, shopping is undergoing a shift, largely driven by smartphones but also enabled by a host of other smart devices:
- From keywords to multi-sensory input
- From ‘walled gardens’ to infinite channels
- From top products to infinite shelf
- From recommendations to perfect personalisation
3. And lastly, the scale, depth and quality of product information is sorely lacking today and essential in enabling the next generation of shopping.”
The idea of founding Indix was based on these three observations, and the AI solutions by the company have these these points underlining its basic functionality.
Future Plans For AI
Indix’s AI-based solutions are being used not only be e-tailers and marketplaces, but also by adtech companies, publishers, market research firms, shipping and logistics companies — in short, any business that contributes to consumer shopping or needs to identify and understand product information.
“Our competitive edge comes from our scale. We have over two billions products in our corpus. More data is equal to faster learning and better algorithms. The large scale of data allows us to develop higher precision algorithms,” said Parthasarathy concluding the chat.
Try deep learning using MATLAB