TFGM AR team receives invoice emails daily in a shared mailbox. Manually extracting customer and invoice details, tracking overdue invoices, and composing professional chasing emails is time-consuming and inconsistent. The 15-day follow-up rule was tracked in spreadsheets with no automation or audit trail.
The AR mailbox becomes an intelligent intake system. Every invoice email is automatically parsed, stored in Dataverse, and tracked. Every overdue invoice triggers a professional chasing email with no human intervention required unless the AI flags uncertainty.
— Rahul Bachina
Triggered by new emails in the TFGM AR shared mailbox. Flow 1 (Email Intake): reads metadata and attachments, calls AI Hub prompt APT_AR_ExtractInvoiceFromEmail (GPT-4) to extract Customer and Invoice fields, upserts to Dataverse, creates Call List and Activity audit records. Flow 2 (SLA Reminder): daily recurrence queries invoices unpaid after 15 days, calls APT_AR_GenerateDebtorEmail to produce professional chasing emails with invoice table. Human-in-the-loop: AR team controls status via Model-driven App Business Process Flow.
- Email-triggered pipeline: from shared inbox to Dataverse to debtor email
- AI extracts Customer and Invoice fields from both email body and PDF attachments
- Auto-generates GDPR-compliant debtor chasing emails grounded in real invoice data
- Human-in-the-loop: Needs Review records require AR team confirmation before sending
- Full audit trail: every email received and sent logged as Dataverse Activity records
- Daily SLA recurrence checks invoices unpaid after 15 days with configurable 7-day follow-up
Want to collaborate or learn more?
Reach out to discuss this project, a partnership, or a new idea.
Get in Touch