Triage Inatake Work Bench

Product Statement

The problem with the current intake process is that it involves multiple manual steps, which are time-consuming and prone to errors. This process requires human intervention at various stages, which can lead to delays and mistakes. Moreover, the process is not scalable, as it requires hiring more resources to manage increasing volumes of requests.

Solution

Automate the intake process by developing a system that can extract email content, apply machine learning algorithms to identify key elements and intents, and store the information in a database. The Phycon app receives email queries and generates workflow objects. The information is delivered to the "Smart app router" via API, where it is received and stored in the database. An hourly scheduler is configured to fetch the records from the table and apply the ML algorithm to extract the key elements and intents from the email content. The extracted information is then stored in the "key words entity table".

Architecture Diagram

Team

Arun Kumar	Kiran Chowdhary	John Crume
Dir Architecture	Principal Engineer	Associate Software Engineer
arun_k_kumar@optum.com	kiran_chowdhary@optum.com	john.crume@optum.com

References

• Triage Power-Bi Dashboard

Github

• Github Link
• TensorFlow Model Files