Parlametric analyzed all customer service chats between the city of Helsingborg and its residents. We identified large differences in negativity between units and what needed to be resolved to increase the positive sentiment. We also ranked most likely questions and best replies based on each unit, and that output was used to build intelligence for a chatbot AI.
Impact
- Identified sentiment, questions and emotions based on specific units
- Resulted in chatbot intelligence with most likely questions and best answer
- Identified negativity trends and how to interject
Difference in sentiment trend between two units
Most common subcategories
Childcare
School/preschool application
Closest school/preschool
Info regarding queue
E-grades
Course certificate
Cancel studies
Apply for courses
Grades
Student loan
Elderly care
Fatherhood certificate
Senior citizens
Meeting point
Retirement home
Nursing home
Translation / interpretation
LSS
Childcare
Elderly care
