Can artificial intelligence accurately measure what you can feel when you look at a new product? It may seem like a stretch, but that’s exactly what Bengaluru-based Entropic Tech is helping its clients achieve. In the words of CEO and co-founder Ranjan Kumar, “they help brands understand their customers’ experience from the perspective of their emotional behavior.”
This decoding of user emotions is expected to be more accurate than traditional surveys. Kumar cites the example of an e-commerce platform that wants to launch a new mobile application. The company can set up a test in which select users go through the new app, and Entropik’s Tech will track their facial expressions and eye movements as they scroll. and record the levels of participation, as well as reactions such as happy, sad or excited.
Kumar says these results in analytics for everything from a mobile experience to an ad or just a piece of content. His key areas of focus are the digital experience, the content experience and the retail experience, he adds.
Entropik’s clients include Procter & Gamble, Flipkart, Tata Consumer Care, Accenture, and ViaCom18, to name a few. A new use case is where educational technology companies are trying to track engagement in online classes.
But how does emotion AI work?
The five-year-old startup has spent its first three years building this technology from scratch. He has also filed 17 patents on the subject and claims to have the second largest repository of emotion data worldwide after MIT’s Media Labs. The company’s models have trained on 30 million emotion data sets and are adding millions more each month.
What Entropik has done is use the latest camera technology to track and capture facial expressions and blink rates in real time. These reactions are then studied by the deep learning model, which deciphers what a certain expression means.
“We can track exactly where you are looking on the screen, as well as. Are you attentive enough, are you engaged enough with the whole experience, etc.,” Kumar explains.
Entropik also relied on brain mapping to create its models, where an EG headset is used to track the neurological pulse of users while watching an advertisement or going through any of the stimulus experiences. These brain wave frequencies also gave the company a good understanding of how to look at behavioral data, especially for accuracy.
“We also look at what kind of facial expressions people have when they watch a certain type of content and then simultaneously we are tracking it from brain wave mapping. We cross-train our facial coding models on that, and that’s how we set up our first one, ”he adds. The company also has a repository of what they call fundamental truth data, based on defined models in behavioral psychology, where a certain facial expression is associated with a certain emotion.
And what is the benefit for customers?
According to the CEO of Entropik, the variety of emotions captured can provide deeper analysis to your customers. For user experience (UX) and user interface (UI) managers, this analysis can help improve the website or mobile experience in a more optimized way. Let’s say Flipkart is looking to have a live application. They do this test. We track second by second as a user goes to make a purchase, where users are happy, where they are excited, what they are looking at or noticing, etc. ”Explains Kumar.
The company also has its own participants and respondents in all countries who have registered on the platform and participate in such tests. All users give their prior consent before joining these trials.
As well as helping to optimize ad spend, this AI technology is also being used to elicit packaging reactions. “In a virtual environment, users can go through an entire shelf, which includes the customer’s bundle and a bunch of competitor’s bundles. We track from eye tracking exactly where they are looking, how much time they spent on a particular brand package compared to the competitors, ”he says.
This technology is not easy to replicate and Kumar has around five competitors around the world. While the technology may be new, it is solving an old problem for brands: how to measure consumer reaction, and accurately.
“Historically, all of these brands have been based on surveys and focus group discussions. In a way, we are betting that 95 percent of consumer behavior is pretty subconscious, only 5 percent is conscious. And that’s where we have a much deeper qualitative analysis. We are able to deliver that value and customers can see a consistent return on investment (ROI), ”he explained.
For Entropik Tech, the next plan is to expand into markets like the United States and Europe, where they see great opportunity for the kind of technology and consumer insights they offer. The company has raised $ 8 million in funding from US-based hedge fund Falcon Edge in September 2020, which they will use to further expand their operations.