Automate the processing of 1,000 requests per week

Written by Martin Bougie

Company

Our client is a company that provides temporary healthcare-related staffing services in various institutions such as hospitals, clinics, prisons, etc.

Challenges

  • The various customer groups send their needs by e-mail. Every week, more than 1,000 e-mails have to be processed.

  • The format of requests is not standardized. Some requests arrive in the form of e-mails written in natural language, others in the form of Word or Excel tables, PDFs, scanned images, etc. 

  • The challenge, therefore, is tousecase1-9nbssimplify and standardize the staffing needs assessment process.

Solutions

  • The solution developed consisted of a mailbox monitoring tool that is able to interpret the data contained in the e-mail and in attached documents.

  • Once interpreted, the essential reservation data is sent to the managers in a consistent, standardized format.

  • Missing data can also be identified to facilitate follow-up.

  • With the help of models based on natural language recognition (NLP) and Large Language Model (LLM) like GPT3.5, the tools developed can understand e-mails and recognize the essential data for extraction.

  • Azure OCR was used to interpret images and PDFs.

  • In the end, a standardized report is sent to the people responsible for the needs assessment.

Results

  • A significant increase in employee productivity by significantly eliminating application assessment time.

  • Increased employee responsiveness to incomplete applications thanks to rapid identification of problem cases

  • Improved customer service thanks to fast, efficient processing of requests.

Martin Bougie
Next

Improving the customer experience