Walmart is developing a facial recognition system to monitor customers at its checkouts for signs of anger and dissatisfation.
The technology uses video cameras at store checkout lines that monitor customers' facial expressions and movements to try and identify varying levels of dissatisfaction, according to a patent filing.
If the system detects an unhappy customer, it will alert employees and order them to report to the register.
It is easier to retain existing customers than acquire new ones through advertising,' the patent filing reads.
'Often, if customer service is inadequate, this fact will not appear in data available to management until many customers have been lost.
'With so much competition, a customer will often simply go elsewhere rather than take the time to make a complaints.'
Walmart will also use the technology to analyze trends in shoppers' purchase behavior over time, according to the filing.
It will link the biometric and emotional data to transaction data in a bid to detect changes in a customers' purchase habits due to dissatisfaction.
'Significant drops or complete absence of customers spending ... may be identified,' according to the patent filing.
The Arkansas-based firm was previously granted a patent for in-store drones that would transport items from one department to another.
It has been suggested that the system would 'greatly improve the customer experience' by saving visitors trips across the massive facilities to fetch items or from having to wait for an employee to return with their desired merchandise.
The patent, entitled 'Method to carry an item within a retail shopping facility' was filed in September 2016 and first reported by Fortune.
Amazon has had its sight set on Walmart for quite some time – as it aims to take its seat as the king of retailing.
However, Walmart appears to be taking a page from the e-commerce giant and utilizing drones to improve customer experience.
The process would first start with a request from a customer, which is sent to a human worker via a 'display screen or as a text message', who then attaches the item to the drone.
The patent highlights a few methods for getting the drone from point A to point B.
A central computer system could be designed to direct the drone to the drop-off point.
Another example suggested using a closed circuit that the drones follow from start to finish.
Regardless of the approach, that patent explains that the drone would fly to the drop-off point with merchandise in tow and release the items once it arrives.