Improved Performance Assessments in Call Centres

In Lithuania, 400 entities provide call centre services, including banks, governmental institutions, insurance agencies, and telemarketing services. Currently, only 1% of these conversations are being evaluated every 2-4 months, which consequently provides inaccurate assessments of the employees performance. These evaluations are also conducted manually, which is a time-consuming practice. 

With this project, Oxus aims to implement the final prototype of the Lithuanian language telephone conversation content quality assessment solution. This includes developing an automatic AI based key word and emotion recognition system which helps to evaluate customers satisfaction of services. By implementing this solution, call centres can:

• Automate and perform quality assessments of all telephone conversations, ensuring compliance with standards;
• Provide almost real-time feedback and insights to improvement for agents conversation quality; 
• Achieve 30% increase in customer loyalty.

Adapting the Prototype to Norwegian Call Centres

IDN will adapt the “OXUS Speech-to-text sentiment analysis” solution to the Norwegian language. This includes the collection of datasets from call centres in Norway as the first activity, followed by preparation and ML/AI algorithm adjustments for the Norwegian language based on what has been built for the Lithuanian language.

IDN will carry out research and experimental development within IDN competence and contribute to the development of a prototype. This includes modification of the prototype for the assessment of the quality of Norwegian conversations.

Co-funded by: