AI and ML-based solutions
R&D. Development. Consulting.
Customer emotion scoring system to determine service satisfaction level. A turnkey solution automatically analyzes service quality and customer emotions from the video stream using ML and CV.
HOME / OUR WORK / Happybox
/ task
Customer satisfaction with the work of front-line managers is very difficult to track on the flow, customers are reluctant to give objective feedback on the service
Scoring of clients' emotions at the moment of service allows to draw conclusions about their satisfaction with the service of an individual manager.
/ Solution
For the solution we used our own development, Happybox — a box solution that automatically analyzes the quality of service and customer emotions based on video data using ML and CV.

Based on the results, the top management received a calculation of the "satisfaction rate" metric for the department as a whole and for employees in particular.
/ technologies
We used our own developments in emotion detection and recognition, as well as the edge computing approach, when the main processing of the video stream was carried out on end devices, which allowed us not to overload the customer’s Internet channels and give instant feedback. For this purpose, neural networks were ported and optimized to work on arm processors. This allowed the product to run on economical single board computers, reducing the cost of integration and the cost of recognition.