Singapore-based ViSenze offers an AI-powered visual commerce platform that enables shoppers to search for products by using image instead of keywords. It counts Urban Outfitters, Meesho, Myntra, Zalora, and Rakuten, among its stable of customers.
Started in 2012, ViSenze says it processes over a billion queries a month from retailers, supporting them in-store and on e-commerce sites. Use cases range from enhanced visual search, product tagging, smart product recommendations, and retail analytics.
Most shoppers know exactly what they are looking for but have a hard time finding it using keywords, and often the recommendations are an overwhelming selection of unrelated products. To solve this, Oliver Tan, Co-Founder & CEO, ViSenze decided to launch his company with the mission of making finding products easier with smart AI.
Working on deep learning and computer vision, the company decided to create a breakthrough technology solution for retailers that is powered by an AI.
This AI starts with understanding what humans see so that their search can be more intuitive without the struggle of keyword guessing.
Tan, who considers himself a ‘bridge between science and real world problems’, questioned the logic behind AI not being used to help solve real problems saying, “If the best science and the technology that we develop aren’t solving real-world problems or pushing the boundaries, then what is the real value of these AI or technologies? AI must advance our daily lives and bring benefits to everyday people. So, in that way, ViSenze is a solution to a real problem.”
Tan, who was speaking at India Fashion Forum 2023, said, “A picture’s worth 1,000 words. But are you going to use 1,000 words to describe the picture? Of course, not! We are all naturally visual people even before we learn languages. I’m just creating solutions that bring everybody back to their instincts – vision. So the native solution is allowing shoppers to shop by taking pictures with their smartphone camera lens or submit images they have captured instead of struggling with difficult search terms to describe what they are looking for.”
ViSenze’s AI Solutions
The company’s underlying AI technology is built on vector-based technology and an inference engine with over 100+ trained AI models that can support a wide range of products from apparel to furniture to other lifestyle products. Most of Visenze’s AI models are trained using deep learning techniques and backed by over a dozen proprietary patents.
“We see technology as an enabler and not an end in itself. It is easy to fall in love with technology and yet not make an impact in the real world. So we are very clear in our roadmap and values to measure the impact it ultimately makes with the AI solutions we create. We see the growing importance of AI technology in a several new areas: personalization, retail analytics, automated inventory planning, and product information management,” Tan stated.
Delivering Customer Satisfaction
On delivering customer satisfaction because of its accuracy, Tan said, “We have applied deep learning techniques as well as real-world data. So when you are able to build robust AI models that are trained on these real-world product feeds and everyday products that people buy, you can get to a very high level of granularity and specificity and yet, be very accurate and relevant at the same time. So the big problem that we fundamentally solve using product visual search, an upper funnel problem, which is dealing with search and discovery.”
“People are searching, but they’re not finding. And the reasons they’re not finding it is either the keywords that they use are inaccurate or not sufficient in depth, or not matching to the taxonomies that merchants have. So many retailers need to solve this by tagging these products with additional search attributes to match with the keywords that you’re using,” he added.
The Indian Perspective
An interesting trend that ViSenze has found in India is the use of screen capture.
“Indians use the image to go and search on a retailer’s site to find the exact product to see whether the prices are cheaper on the other side. So we realized that there were a lot of screen captures from Instagram and from malls. All the social platforms are used to compare products and prices. Price comparison is a very strong use case in India,” Tan noted.
Image search in the Indian e-commerce industry registered a 180% growth in 2022, according to a new report by ViSenze. The report added that image search was used 1.4 billion times last year on e-commerce websites and apps across the country.
Fashion and apparel were the most searched items, making up 72% of all image searches,
while footwear, jewellery, and ethnicwear also saw a significant increase in product searches, accounting for 44% of all searches, said the report.
The Indian Fashion Market
“What I realize about India is, number one, that’s a massive country with a massive online shopper base, which is set to overtake the US as the world’s second largest shopper base in two years’ time. Number two, the pandemic has pretty much accelerated a lot of online traffic. India has high smartphone penetration. Number three, fashion is a big market in India because you have a pretty large young population. These young shoppers have adopted and adapted to not just the online and offline customer journey experience; they have also adapted to the new tools that big leading retailers offer. So, this has overall lifted the market in terms of adoption,” said Tan.
He added that ViSenze, which counts Myntra, Ajio and Meesho among its Indian client list, is not here to ride the wave but to help.
“India has a vernacular problem; there’s no common universal language like Mandarin in China. Of the top ten languages spoken around the world, I realized that three are from India. I was very amazed-how could one country have three major languages? It is not just this; there is also diversity at the local level. Some of these vernacular languages have no keyword equivalent,” Tan pointed out.
Tan said that doing business in India requires a lot of adaptation and understanding of the cultural differences. But really understanding the root of the problem is the biggest challenge retailers face here.
“Coming from a small country like Singapore, we probably have solved half of it because it’s so much more manageable, and you get it done very quickly. Here we’re dealing with a massive scale of problems. But you have a lot of untapped potential in tier 2, tier 3, and tier 4 cities, so to speak. Doing business in India requires me to have a deeper conversation. What are you trying to solve? What keeps you awake at night? What are your priorities? How can I help you? I like to get the fundamentals in understanding how
to help our clients,” he explained.
The Year 2023
Tan feels that we have come back to a more normal state of business post-Covid but not without some fundamental shifts. So this, to him, means we’ve got to seize the opportunities and start reinitiating deeper sets of conversations with retailers and brands from around the world.
He asserted that omnichannel is not just a buzzword; it’s also retail culture in addressing these fundamental shifts, such as a changed customer journey. So as an AI company, he says we cannot ignore the realities of customer’s journey, which is being impacted by Covid and have emerged in the post-pandemic era in more ways than one.
“A typical customer may not start his or her journey on Google. It may start on Instagram, Youtube or even in the store. But the question is, how does the retailer, whose products the average shopper is looking for, show the most desired products to the shopper? And Where will they make the purchase?” he asked.
The answer lies in truly appreciating the changed customer journey in a phygital world, and making it as simple and as easy as possible for the customer to make a purchase decision.