Atul Rai, CEO & Co-Founder, Staqu, on how AI can be used as a tool to accelerate the businesses

Atul Rai

Staqu, a Gurgaon-based Artificial Intelligence (AI) startup, launched last year, has been working for mainstreaming Artificial Intelligence. It allows users to search for the desired products by uploading same or similar images. The technology was explored by Paytm and YepMe and an OEM, Karbonn, Lava and Panasonic. Staqu further launched Fashin, an iOS and Android app that extracts fashion trends from running YouTube videos. Atul Rai, CEO & Co-founder, Staqu shared his views on the usage of AI as the tool to accelerate the businesses with Haider Ali Khan that how it can be used especially by the Indian startups to stand up and compete with their global counterparts.

Why Staqu, when the consumers have Google to do image search also? How does Staqu create a niche for itself in this ecosystem?

Google image search is more like a generic search engine and is not meant for searching a particular dress/fashion product. For example, if I will upload an image of a dress worn by a girl, in Google image search, it may start searching similar pictures of the girl instead of the dress. Also, neither Google provides APIs to e-commerce like we at Staqu do with our VGrep API, nor does it have any such business model for e-commerce companies so far. If I were Google, I will not limit myself to a limited category of fashion because Google is used and known for generic search support not domain specific search.

Also for unstructured domain developing an AI solution would need focused approach as well as data (even for subdomains like ethnic, western etc.) There is a vast difference between developing AI for general image recognition and developing domain specific image recognition/ understanding API.

Furthermore, unstructured e-commerce sector is not uniform or singular throughout the world. Various e-commerce sites existing and operating in different countries have a different set of aspirations and requirement. For instance, an AI trained on western data may wrongly recognize Lahenga as a dress and Nehru suit as mandarin collar shirt in India.

Having a uniform API powered solution will never work for such unstructured market. Hence, one needs to customize the solution as per the requirements and that is where a dedicated research, engineering and business team would be needed to match these aspirations. While for the tech Giants such as Google, this would be just a side product, for startups like Staqu, image search for ecommerce is the main product and thus, our focus will be clearer and more focused when compared to the same.

We still lag behind when it comes to using AI. How can Staqu act as a disruptor in this genre?

I somewhere don’t agree with that statement. Most of us are using Facebook, Google, Snapchat, Prisma in our daily life and all these products are majorly powered by Artificial intelligence. Besides, there is another misconception amongst people that make them treat it as an end product.  We need to understand that AI is a technology that, if wrapped around different products, will provide superexciting, smart and intuitive results. And we have already experienced the same in the products or platforms mentioned above.

Which are the platforms is it available at? Do we need a separate app for everything like emotions, image search or fashion?

I believe yes, we would need a separate app for separate domains. As I said AI is a technology and not a product and it can be amalgamated with different products in various domains. Hence, while a fashion company would need image search capabilities for searching apparels, a medical imaging company, on the other hand, would use the same image search technology to identify tumors in the given image. The same goes for emotions recognition and other domains.

What are the pressing problems which e-commerce is facing today?

In general, e-commerce companies are facing two major pressing problems. Firstly, logistics and secondly, cataloguing. We at Staqu are working to solve the latter one. When I talk about cataloguing, it includes the consumer facing search and description of products, along with Trend analysis, strong recommendation engine etc. Most of these works require manual human interference, which in turn drastically increase the cost and time. We at Staqu are working to automate most of these tasks to make the processes more cost- and time-efficient, with human level accuracy.

How can OEM leverage the benefits of AI from Staqu?

So far, most of the OEMs are competing with each other on the basis of hardware, with only marginal profit. However, a mobile phone is not just hardware but a collection of so many software components. Infusing these software features with AI will not only benefit the end consumers but will also generate a new revenue stream for the mobile companies.

For example, we can make image search part of the camera itself. Users would then have the freedom to search whatever products they click, perhaps buying the same in real-time. It will benefit the user but at the same time, will open a new revenue stream for OEMs, Ad Tech and more.

How crucial AI is for the e-commerce and m-commerce?

AI is just not crucial to e-commerce; it is soon going to become crucial for every domain that wishes to make the existing processes efficient and cost effective. I consider AI as the next revolution, post the advent of machines.

In e-commerce, for instance, products are needed to be tagged so that an end consumer can search it through textual query. In the same pursuit, a minimum of Rs 5-15 is spent by e-commerce companies on every product, along with additional costs for infrastructure and the time consumed in tagging each of the products. AI can make the same process rather efficient and optimized. Hence, while a single curator may tag a maximum of 700 products every day, AI can do the same within seconds.

What can Indian startups learn from Amazon when it comes to usage of AI?

Indian startups may not possess the same wherewithal as Amazon, when it comes to exploring the depths of AI, machine learning and automation. However, the success of the ecommerce sites can be replicated or better yet, over passed by the ambitious Indian brands. Amazon provides Indian startups with a potential learning model, especially with the fact that it has never shied away from deploying technology in order to achieve business objectives. Indian startups have their own unique challenges at this moment; however, if we strive towards the direction of technology and innovation, the success is going to be ours in the long term.

Leave a reply

Your email address will not be published. Required fields are marked *