Next time you surf for that handbag on an e-commerce site, shortlist the product, smile while looking at it and move on without buying. You may get an offer to entice you to buy it soon enough. Thanks to device cameras, consumer targeting based on facial expressions are soon going to become a reality, with companies such as Singapore-based EnableX partnering with offline and online retailers in India.
“We are talking to some fairly big e-commerce companies both in India and overseas, where they are looking at enhancing the customer experience and personalising their shopping experience,” said Pankaj Gupta, Chief Executive Officer and founder, EnableX.io.
While he did not share names, one of the companies that have signed on is a Chennai-based e-commerce enabler.
Other companies also working in this domain include Entropik Tech, which uses emotion AI, brainwave mapping, facial coding and eye-tracking to decipher the cognitive responses of customers.
Facial recognition and emotion analysis
EnableX provides business intelligence services to its partners through facial recognition and emotion analysis with its solution, called FaceAI. The idea is to capture customers’ emotions and help etailers offer them the right products.
The company claims that its technology can detect a shopper's gender, age, emotions, among other things, in real-time. This, in turn, helps target them with relevant ads and promotions.
Using FaceAI, businesses and developers can integrate emotional intelligence into any application, website and digital displays.
When it is deployed, this could prove to be a game-changing technology in the Indian e-commerce space, which is projected to be worth $200 billion by 2026.Currently, there isn’t much of a look-and-feel about the whole online commerce process. Through such tools, especially by virtue of the camera, etailers will be able to know a lot more about the audience though all this is still a long way away, said experts.
“Startups are evolving. There are a bunch of innovations happening around this space but we are still in the very early stage of the hype cycle. It’s all about if the application can figure out if the viewer is in the right sentiment, in a mood to buy or not. Then they can offer them a discount or target them right way,” said Sandip Kumar Panda, founder of cloud-based SaaS-services provider Instasafe.
While all this looks promising, such a service has also raised huge privacy concerns across the globe.
"Now every good thing has got an opposite. It is all about user awareness. From a privacy perspective, this is obviously crazy, but overall, AI will even out in those terms. When I say even out, I mean cybersecurity technology is also becoming rapidly AI prone. So, for every action there will be a reaction,” Panda added.
To address this concern, Gupta claims the company follows GDPR compliance (General Data Protection Regulation — an EU norm to protect personal data and privacy). All the AI processing is done on the device with no personal data, images or videos shared with the centralised EnableX.io server.
But how will the app know if a customer raised her eyebrows in shock, criticism or excitement? That can only be ascertained if the company has a verifiable statistical sample. For example, many customers may frown in surprise. So, only if the app can statistically correlate why half a million users are frowning when they’re surprised “will we have a statistical average that this reaction is valid.” said Panda.
Some people may giggle and not frown when they are surprised, so that's a false negative. In the false negative, the algorithms go for a toss, Panda added.
Therefore, any artificial intelligence is only possible if there is a lot of data sampling.Is it possible to collect the relevant data and still maintain privacy? To some extent, yes, says Panda. For instance, he explained, such companies can make a graph out of a customer's face and reaction, which shows whether or not she is frowning. But this will not store the customer's name or personal data. Then the collected data is randomised to achieve the desired results.