Privacy-first emotion AI: How Wayvee Analytics tracks customer engagement without creeping you out

As a journalist with many years of experience writing about IoT, I've seen its innocuous embedding in every sector. A key example is retail: inventory management, supply chain optimisation, point of sale, shopping carts, and customer behaviour analytics.

Customer behaviour analytics has traditionally relied on video-based foot-traffic monitoring and heat maps to help retailers understand customer movement patterns throughout a store so that they can improve store layout, product placement, and, ultimately, the customer experience. 

However, surprise, surprise, the use of IoT in retailtech also comes with privacy concerns, especially in the case of retailers which use advanced tracking methods like facial recognition, high-tech cameras, and sensors that can capture voices and geo-locations. Further, many retailers (outside of Europe with its GDPR protection) collect data without obtaining explicit consent from customers, violating privacy laws and regulations.´

However, a retailtech startup with roots in Cyprus and Poland, Wayvee Analytics, has found a novel way to track customer engagement in brick-and-mortar stores. Its foundation is old-school tech: radio frequency (RF) waves that, combined with proprietary AI algorithms, analyse emotional responses in a privacy-first way. 

I spoke to co-founder and CPO Alex Ovcharov to learn more. With a background in neurobiology that includes Business Development Manager at MediaCom, Product Director at Shazam's Eastern European division, and Deputy CEO Sensemitter (a fascinating company in its own right focused on emotion AI analysis in gaming), Ovcharov brings deep technical experience. 

Old tech to solve the privacy-problem

In-store tracking is nothing new, and besides the problem of privacy, it faces numerous technical challenges. Ovcharov detailed using cameras to track emotions:

"For it to work, you need a person to be very still, and for precise measurement, you need at least 15 cameras for one shelf. It's not only expensive but unviable. 

We once installed 100 cameras in a bank office to evaluate the whole space. The employees told me it was a great experience, but we thought, "Never again." 

Ovcharov came across the idea of radio wave tracking of the human body 

"As a neurobiologist, I figured out that if the radio wave can meticulously measure the change in the position of a human body through the walls, then it definitely can measure the breathing because it's also a movement as well as the heart rate, both of which are easy to translate into an emotional response."

Installed in various retail settings like shelves, displays, in-store media, and point-of-sale (POS), the Wayvee Sensor uses RF waves, which contain data on physiological responses, such as breathing rate, heart rate variability (HRV), and body gestures. Wayvee's AI algorithm then analyses these responses through a trained neural network to evaluate customers' emotional states and convert them into customer satisfaction (C-SAT) scores.

This results in actionable recommendations to improve in-store outcomes. For example, retailers can adjust product placement or update rotating content to improve outcomes and enhance the overall customer experience.

How Wayvee quantifies emotions for business insights

When I first encountered emotional tracking in wearables—bear with me, it was a long time ago—my question was always around the ability of data analytics to distinguish between heightened emotions such as excitement, fear, and joy. Otherwise, it's nothing more than an old-school mood ring. 

According to Ovcharov, traditional emotion theories, such as Ekman's, emphasise subjective emotional experiences and rely heavily on facial expressions as indicators of emotion. 

However, the subjective nature of these interpretations and the limitations of facial expression analysis in accurately reflecting internal emotional states hindered the development of reliable emotion-tracking technologies.

Ovcharov explained:

"To address these challenges, we've adopted a behaviour-based approach. This approach focuses on objectively measurable physiological signals, such as arousal levels (the intensity of the emotional response) and valence (the positive or negative quality of the emotional experience).

We also recognise the crucial role of engagement, or the level of involvement in a decision-making process, in understanding customer behaviour."

This data-driven approach identifies the strong correlations between physiological arousal and key business metrics such as sales and customer satisfaction.

While the startup is still in its early stages, it has enormous potential as stores look for more solutions to increase customer loyalty and satisfaction. 

And if you're reading this thinking that most people shop online these days, you'd be wrong. According to Ovcharov, 86 per cent of grocery retail shopping occurs in-store. 

"Many consumers have established routines for in-person grocery shopping, making it a familiar and comfortable part of their daily or weekly lives.

Further, people live in close proximity to a grocery store, making it convenient to shop in person. It's an integral habit." 

One of the most significant growth areas for Wayvee's reach will be the use of dynamic pricing on electronic shelf labels. These are used to enable retailers to adjust prices to match market conditions or competitor offerings quickly and when combined with customers' emotional data, Ovcharov predicts:

"This will be the most AI-powered thing you can imagine because this is not just the synthetic data; this is responding to real-time human behavioural data.

Bricks-and-mortar stores suffer from very thin margins. This has the potential to increase customer satisfaction, which increases in-store revenue growth and customer retention. 

The potential to optimise in-store advertising and promotions

While Wayvee's customers are currently brick-and-mortar stores, there's huge potential in the future to extend their focus to the brands that inhabit the shop shelves. 

"Currently, we are in a blue ocean market. We are starting with a solution for stores because it's more scalable in terms of business, but in terms of brands, I see the impact here because brands would like to monitor their goods in the in-store environment.

Further, our device is very appealing because it doesn't breach any privacy regulations. Any  brand could deploy the tech to determine optimal positioning and in-store marketing."

There's also scope to extend Wayvee to beacons, which some retailers use to send targeted promotions and discounts to customers' smartphones as they move through the store. 

And then, Wayvee emotional response analytics would also work with in-store advertising. 

Another practical use for Wayvee is to predict shoplifting. Ovcharov admits this is morally a thin line but stresses, "Such a use case is not about tracking a person around a store but rather identifying peak points where a high level of arousal signifies they may need more security. 

This can actually prevent people from shoplifting in the first place through target hardening. 

While still in its early stages, Wayvee has the potential to transform in-store analytics by providing valuable insights into customer behaviour without compromising privacy. it holds significant promise for both retailers and brands, enabling them to optimise their in-store presence and enhance the overall customer experience in a data-driven yet ethical manner.

Lead Image: Wayvee Analytics. Photo: uncredited.