UC develops innovative solution for guiding call center agents
The main goal of the Flowance project was to develop support mechanisms for contact centre agents so that they can be more work efficient.
English version: Diana Taborda
A research team from the Department of Informatics Engineering (DEI), Faculty of Sciences and Technology, University of Coimbra (FCTUC), has developed an innovative platform that aims to provide an efficient real-time guidance to human agents in call centers.
The project “Flowance - Automatic Dialog Flow Extraction and Guidance”, co-financed by the European Regional Development Fund (FEDER), is led by Talkdesk (a world leader company on technological solutions for call-centers), in a partnership with the Centre for Informatics and Systems of the University of Coimbra (CISUC).
"The main contribution of CISUC in this project was the development of a platform to identify the most frequent dialog flows from a set of dialogs, greeting, request for information, response, clarification) and the probabilistic transitions between the different states of the flow”, explains Hugo Gonçalo Oliveira, CISUC researcher and Professor at DEI, further adding that “we had to try different ways of representing, classifying and grouping dialog utterances, explore different ways of interpreting and labelling the flow states and, more recently, new ways of visualizing the extracted flows".
According to the CISUC researcher, we expect these flows to offer an alternative way to analyse interactions and, when combined with their visualization - like graphs -, to identify patterns in communication, such as unexpected situations, requests for information, and the most common responses.
Although the Flowence project has just ended, the team continues to work on this solution under the project “NextGenAI - Center For Responsible AI”.
In the new project, the focus is still on using these methods to extract flows and help explain the functioning of chatbots that are not based on predefined flows, such as ChatGPT, thus contributing to a greater confidence in their use. The project proposes to prevent call-center service degradation and customer dissatisfaction through the use of these advanced chatbots, taking advantage of technologies that can make a human’s job more efficient without losing human contact.
The CISUC team consists of researchers Hugo Gonçalo Oliveira, Catarina Silva, Ana Alves and Nuno Lourenço, and PhD student Patrícia Ferreira.