[Note: This article is part of The Junction Series. We will be covering the RetailTech and DeepTech sectors in detail at The Junction 2017 in Jaipur. Learn more about The Junction here!]
One of the new Lord of the Geeks, Elon Musk believes that the world will one day be taken over by robots who will rebel against the humans and make them their slaves. A scenario that has already been acted out in movies such as “Terminator,” “iRobot,” and “Eagle Eye” to name a few.
The plot usually traced these robots having an evolved form of intelligence, created by humans themselves, called Artificial Intelligence (AI) that they used to plan their global domination and rebellion. And this technological explosion happened gradually enough for no one to be the wiser. While this fictional doomsday scenario maybe a bit farfetched (and depressing!) at the moment, it would be foolhardy to disregard the many ways AI is pervading our ordinary Indian world. India boasts of 250 Mn smartphone users at the moment and is moving towards a digital-first economy.
But there are hardly any frontrunners in a business that is expected to reach about $5.05 Bn by 2020, showing a staggering 53% overall growth, and this is why it is an area that is worth exploring.
This notion convinced AI-based Staqu’s co-founder Atul Rai to study AI in college, which laid the foundation for him eventually founding Staqu – an AI-based startup that provides products and solutions for OEMs, B2B, and B2C platforms. But to get from student to Jedi, he did extensive AI-based research in 2009-10, to the extent of publishing a research paper in France on the various ways AI could possibly help the physically disabled segment of society through gesture recognition, facial recognition, and more.
From An Academic To An Entrepreneur: The Staqu Origin Story
“I have always been an academic,” begins Atul. “I did all this research during my graduate days and then went on to complete my Master’s from the University of Manchester which is famous for AI testing,” he shares.
Doing extensive research made one fact startlingly clear to Atul – 70% of data shared on the Internet consists of images.
“Videos are just catching on, especially in India with our slower data connections, but back then in 2010, it was abundantly clear that people preferred looking at images and searching for them in order to make lifestyle-related decisions and people were not really providing a comprehensive package that would collate all this data and research that existed and use it to build out their business,” he says.
Atul continued working on the research for a few more years, his forte, solidifying his theories and figuring out how tech-focussed AI could be interpreted for regular customers. He saw that the market was uber-nascent in an emerging country like India. His idea was sound and based on extensive research. He also realised something else.
“To do research is not enough to sell an idea. You need a product, a business.” But Atul was undeterred and he pressed on in search of the perfect combination of product and team to see his idea come to fruition.
This was around the time that Atul met his co-founders Chetan Rexwal, Pankaj Kumar Sharma, and Anurag Saini in 2015. “Chetan and Anurag were expert coders,” he says. “They had experience building out mobile-based apps or native apps which could be integrated directly into the OS of a mobile and deployed instantly. No hassle, no work. And thus all of my years of research finally found an outlet as it turned into product development,” he adds.
“Our team fit was complete with Pankaj joining in, to takeover server development, as a business that is going to sift through copious amounts of data needs a strong network infrastructure to begin with.”
Staqu was formally launched in 2015 with a suite of AI-powered operations including bi-directional image understanding, an innovative image-to-image matching system, simplifying image search, automated meta-tag generation and real-time product recommendations.
Using AI To Understand, Automate Consumer Behaviour
According to Atul, Staqu works on two levels – B2B and B2C2B. On the B2B front, it works almost exclusively with ecommerce companies in areas such as automated meta-tagging, image search made easier, visual recommendation, and complete look recommendation.
“Ecommerce companies depend mostly on images to effectively showcase their product SKUs and run their business. I would even go so far as to say that ecommerce sector depends 70%-80% on images for a customer to get interested in them and complete a transaction,” Atul says.
“Now, say you’re a customer who wants to buy a Blue, Lucknowi kurti but typing all of this takes ages and there is no guarantee it will give you the exact blue, Lucknowi kurti you wanted in the first place. You basically cannot describe every single aspect of a product via text, some of it has to be intuitive and/or automated.”
So, what Staqu did was build an API that was integrated in the database of the ecommerce site for immediate deployment – Image Search – wherein you take a picture of the product you’re searching for, upload it on the site and the search will bring results within seconds.”
Atul also reveals that this native app and API has been used on emerging ecommerce and retail platforms such as YepMe, Roposo, and TradeIndia among others.
He mentions two other pain points of ecommerce marketing- meta keyword-tagging and trend analysis which are also solved by using Staqu products. For meta keyword-tagging, the idea is to eliminate human error by automating the process and tightening it for SEO purposes as, so far, the curation for keywords is done by people.
For trend analysis, Atul shares,
“At the end of the day, ecommerce companies are looking for new ways to engage with their customers and appeal to their wallets by giving product recommendations. This process is time-consuming when deployed by humans, but automate it using social media profiles of the users and you have a whole new revenue stream ready to be tapped into.”
Using AI To Influence Consumer Behaviour Through OEM
The other product that Staqu has launched is for OEM (original equipment manufacturers) seeking to integrate its solutions at the manufacturing level.
Atul shares the use case of Karbonn Mobiles to elaborate on this:
“Karbonn Mobiles carry in-phone the Fashion Eye software which basically means that the phone’s camera works as an image search and tagging platform. We then take this search a step further and have tied up with ecommerce platforms where you can find the items you clicked and searched for on them, or the closest recommendations to them. It is a smart technology integration without the need to download any other app on the mobile.”
While the idea certainly seems cool and futuristic, an interesting take on Point-Click-n-Buy, the startup was not short on challenges convincing partner clients to take it on.
“The partnership with Karbonn took nearly 3-4 months and many meetings to resolve their queries and doubts regarding the effectiveness of the platform and product. AI is still so nascent in India, it was a challenge to get people to accept that the platform and product would indeed work. We did proto-testing, product-testing, beta-testing before the agreement was signed,” shares Atul.
But interestingly, he also claims that once the first partner came onboard, things became a lot smoother. And the startup got nine more partners within six months of launching.
The startup’s next level innovation comes in the form of Fashin.ai, an iOS app that was launched in October 2016 and has reportedly garnered more than 1,000 downloads. According to Atul, Fashin will be of particular interest to Bollywood lovers who are also obsessed with fashion trends and will be comfortable opening their wallets for purchasing the same.
It is an AI-powered video curation engine where “you feed the video of the song/scene of your choice and it will extract the fashion in the video along with recommendations on where to buy the same.”
“Fashin is currently our customer-facing product on the B2C end but we are also rolling it out with a few big, B2B partners,” he adds.
Business Model, Traction
Fashin was launched on October 12, 2016 on the iOS platform as it was a preferred platform for luxury users. And Atul claims that the app has seen more than 1K downloads in the intervening time, organically.
Staqu’s business partners range from OEMs like Panasonic, Intex, Lava and the aforementioned Karbonn.
The business model works on a license-based subscription for the number of products/features used by clients. For OEMs it is on a commission-basis that is charged by the affiliate partners such as Flipkart, YepMe etc.
And then, there is always lead generation via the use of native app Fashion Eye – “Imagine taking a picture of the outfit you love and finding it on one particular platform and then you find a bunch of other things you love there too and turn into a paying customer. That’s got transactional value,” says Atul.
But when asked about the unit economics of the startup, Atul says, “We believe in building a business, not unit economics. AI is very young in India more so than globally and getting people to believe in the product and building it matters more anything.”
Even so, Atul claims that the company will attain operational profitability in the next two quarters and a M-o-M growth of 30%, far.
“With respect to the B2B arm, we are not marketing at all, as it depends on the class and strength of our partners. But we do believe that the market is set to grow with more homegrown manufacturers coming into play across the board in India,” he says.
It is also a testament to the quality of the product that Staqu has received Seed funding from IAN.
“Developing Tech To Bring Ecommerce Into Every Mobile Phone”
The year-old startup competes with contenders such as Niki.ai (which has recently raised funding from Ratan Tata and Unilazer Ventures), Active.ai (which received $3Mn recently), Fluid AI – which is also looking to change the way customers react in shopping situations via helpful assistant touch screens. But Staqu’s plan to deal with competition is to boost product and technology extensively in order to bring ecommerce in to the umbrella of mobile utility by default.
“Currently, all phone users use like one-two apps for shopping, one-two for news, lifestyle etc. We would like to change this dynamic by offering a utility tool/product like Fasion Eye and Fashin that lets the customer choose from a larger variety of products and, at the same time, give ecommerce players an edge in terms of tech and lead-generation,” says Atul.
It is also constantly looking to innovate and improve on AI research, with academic Atul at the helm to automate interaction and change the way the tech itself is used. For this purpose, the startup has tied up with IIIT Delhi and is open to publishing papers for academic research on the effects of AI on tech verticals in India, among other things.
“When it comes to our B2C models and Fashin in particular, which should be rolled out for Android this month, marketing is required. To build awareness, to engage customers and so much more. And that’s where the money will be used,” he shares.
For the 19-member team from Gurugram, influencing the world of fashion through AI seems just like the thing to do.
Editor’s Note
AI, as a technology, is at par with Visual Reality/Augmented Reality, machine learning and natural language processing, which are also subsets of AI. But the use of AI in enterprises are mainly related to automation and intelligence-gathering. The global AI market was valued at just about $419 Mn in 2014. Now, it is projected to reach almost 20x by 2020. The scope for industries to turn to AI for solutions and insight is immense and immediate, since it is one of the few frontiers left to truly conquer.
In India, the AI party has begun late as industries and verticals have focussed on software, hardware, analytics, and even IoT so far. AI as tech that merges hardware and software is a whole new ballgame that requires extensive research before it can be executed into a viable product.
In the rarefied world of fashion, Staqu might have struck a nerve with its customer-friendly AI products, but whether it can sustain the momentum and fend off competition from the more established players is something that only time will tell.
