Techblog
How ViSenze brought Artificial Intelligence (AI) to the e-commerce market
Siri, Alexa and self-driving cars – these are examples of artificial intelligence at work. Singaporean startup ViSenze uses the technology to run visual searches, for e-commerce players such as ASOS and Rakuten.
The era of online shopping is upon us. While the physical store has its charm, increasingly shoppers are checking out products and buying them online.
Recent research by Google and state investor Temasek Holdings predicts that the value of e-commerce in Southeast Asia will rise from US$5.5 billion in 2015 to US$88 billion by 2025.
Against this backdrop, being innovative in the e-commerce space makes perfect sense. Founded in 2012, Singaporean startup ViSenze does just that. It uses artificial intelligence (AI) technology – widely used in the transport, service and military sectors – to boost sales on e-commerce websites.
For years, customers have used keywords to search for items. The ViSenze technology instead taps AI, which allows customers to upload a photograph or video instead. The e-commerce website in question would then automatically come up with similar offers from its inventory.
This transforms the customer journey “into a visually exciting experience that eases search, personalisation and engagement,” says Mr Oliver Tan, Chief Executive Officer of ViSenze.
“ViSenze’s technology provides consumers with an easier way to find products at their exact moments of inspiration, while also opening doors for brands to ensure their products are discovered across all channels by consumers.”
The effectiveness of the technology rests on how well it can study an image and go on to curate results for users. “Our deep learning and computer vision technology processes information that lies within images and videos to extract intelligence that text-based search engines are unable to,” says Mr Tan.
“When a consumer uses ViSenze for visual search, our algorithms recommend products based on colour, shape, pattern and style, and can identify macro and micro trends.”
AI's building blocks
Still, there is no one-size-fits-all approach to AI. Constant improvement is crucial, especially with the emergence of new technologies. Other challenges that AI offerings face include data acquisition, speed of training and scalability.
“Our machines are self-taught on how to recognise objects and attributes based on examples provided by our customers, and … are constantly being improved,” he says.
Founded in 2012, ViSenze is the result of a programme that brought together National University of Singapore computer scientists and web entrepreneurs, to solve real problems through the innovative application of NUS technology.
It started small, but currently has more than 50 full-time employees. They work from the Singapore headquarters, and global offices located in San Francisco, Dublin, Beijing and New Delhi.
Key to the company's expanding operations is the research and development team. According to Mr Tan, they are currently testing and developing options including video commerce, social commerce and visual analytics.
“All of these improvements will ultimately help retailers solve consumer search problems, recommendations, personalisation, and help better understand consumer intent and behaviour,” he says.
The company also plans to “double down” on key markets such as the US and UK, as well as secondary markets such as South Korea and Japan.
ViSenze has clients from the jewellery, interior design and intellectual property sectors, but most come from the realm of fashion e-commerce. Popular websites like Zalora, ASOS and Rakuten have benefited from the company’s technology.
Shopping for the future
Visual search is gradually becoming second nature to today's customers and leading retailers have been implementing these tools on their e-commerce websites. Platforms such as Pinterest and Google have also developed visual search tools of their own. “It's clearly a tool that more users want and expect,” says Mr Tan.
“As the web becomes increasingly more visual, customers want the ability to engage with brands and products in a way that’s native to the visual content they’re already consuming.”
Making the most of e-commerce with AI is not a farfetched option for small business owners either. Technology like ViSenze can open doors for retailers and brands that typically don't have significant research and development budgets, or the resources to build AI tools.
“This is where we play a significant role in the market as we can drive intelligent product recommendations for retailers because of our efficient pre-training of algorithmic models based on real data,” says Mr Tan.
“As opportunities expand in the e-commerce market, AI's barriers to entry in retail will dramatically decrease. Internet and traditional retailers who decide not to adopt AI will quickly find themselves outmatched by those who do.”
Mr Tan has previously summarised his company’s innovative use of AI: “If a picture is worth a thousand words, do we need a thousand words to describe the picture? Why can’t we use the picture to search for other pictures?”
Summary
-
While the physical store has its charm, Southeast Asian shoppers are increasingly making their purchases on e-commerce sites
-
Singaporean startup ViSenze uses AI – widely used in the transport, service and military sectors – to boost sales on e-commerce websites.
-
A customer can search for products by uploading a photograph or video. Algorithms recommend products based on colour, shape, pattern and style
-
Visual search is becoming second nature to today's customers. Major retails and platforms like Pinterest and Google have implemented visual search tools of their own