Streamlining Utility Bills/Invoice Processing: Process Automation

Email Monitoring: We leveraged Power Automate to monitor designated email inboxes for new utility bill arrivals. This ensured that bills were promptly detected and processed without manual intervention.

Classifier Component: A classifier component was developed to categorize utility bills into categories such as Gas, Electric, and Water. This component moved the bills to respective processing folders stored in Azure Blob Storage, ensuring efficient organization.

Attachment Handling: We developed a Python component to efficiently split attachments, especially consolidated PDFs, into individual bills. This step was crucial for seamless downstream processing and ensured that each bill could be processed independently.

Azure Document Intelligence: We utilized Azure’s Document Intelligence service to analyze and extract required data elements from the invoice PDFs. While this service provided valuable insights, additional post-processing was required to extract missing values and perform cleansing and normalization activities.

Python-based Post-Processing Component: To address the limitations of Azure Document Intelligence, we developed a custom post-processing component. This component was responsible for extracting missing values, performing data cleansing and normalization, and preparing the data for further processing.

Output Preparation: We developed a Python-based component to map the extracted, calculated, and normalized values into the final output schema. This component generated CSV outputs in the desired format, ready for use by downstream systems.

Automated Validation Alert Mechanism: Finally, we developed a Python-based validation component to validate the outcome of each step and have enabled automated email alerts on validation failures or any other errors/exceptions.