IMPLEMENTATION AND TESTING

​

Problem

Centrica People Services (CPS)- the call

centre that handles employee problems is not very efficient - only 50% of calls are resolved. HR asked us to help them identify the most valuable problems to fix and are keen to know how artificial intelligence could possibly reduce the need for CPS

​

Approach

Run 2 years of CPS call log data through a custom designed Topic Modelling Tool to identify most pressing employee problems. This was filtered through call duration and resolution times. This allowed us to pinpoint which problems to address which offer the most value to employees and the business

​

Heuristic analysis of the employee online help system was conducted. Call listening and an ethnographic review of the help system was conducted. The user journey was also mapped. Quantitative research was run with whole company to understand employee sentiment and appetite around help systems and conversational interfaces.

​

Findings were collated into a report and presented to Human Resources. Recommendations were made for online help site improvements and a potential appetite for a conversational interface.

​

Outcome

Based on our research, a conversational interface has been commissioned by HR. We are now working with the AI team on iterative sprints to build, test and refine a desktop based conversational employee assistant