challenge
The customer is a leading company in Switzerland for engineering services in building and civil engineering. Every month, the company receives 300 to 500 inquiries from customers, partners, suppliers or applicants via e-mail or via the homepage. Sorting and dispatching to the responsible employees caused a great deal of effort in the back office, sometimes incorrect allocations and thus delays in processing. The goal was to automate the process of email categorization and forwarding to reduce the manual workload of employees and increase their overall operational efficiency
solution
- Development of AI-based algorithms for email classification and forwarding
- Use of transfer learning and use of a pre-trained transformer-based language model (huggingFace)
- Provision of the solution based on Microsoft Azure and integration into the customer's existing system environment
- Continuous optimization of the AI solution
Business benefits
- In 98% of cases, an incoming email is correctly classified and assigned to the right employee
- Manual activity was largely automated
- Saving time in the back office
- Significant acceleration of the processing process for customer inquiries
