Inbox Butler
Two automation pipelines that replaced manual inbox monitoring and data entry for order processing.
- Status
- Validated on live orders
- When
- 2026
- Role
- Built at work
- Area
- Professional work
In plain English
Inbox Butler replaced manual inbox monitoring, data extraction and data entry for a legal and commercial services provider. One pipeline reads lien search results and writes the details to the downstream system. The other watches a shared order mailbox, classifies each order and enters it either straight into a database or, through browser automation, into a workflow system that has no API.
How it works
- Parse UCC lien search results, extract filing, lien, debtor and creditor details, validate them and write them downstream.
- Monitor a shared order mailbox through the Microsoft Graph API and classify each incoming order.
- Extract the order's details from the email and its attachments.
- Write the order straight to the order-management database, or enter it with headless Playwright in a workflow system that has no API.
- Record receipt, classification, extraction, chosen path and completion for every order in a shared audit log.
Security and outcome
Security was built in: credential-free sign-in with Azure Managed Identity, central secrets management, mailbox access scoped to the one shared mailbox, encrypted traffic, and separate dev and production environments with pull-request approval gates. Both pipelines went through end-to-end testing and subject-matter-expert (SME) validation on live orders, and SME testing estimated about a 70% improvement in operational efficiency for the affected workflows.