AI Invoice Billing Routing Agent

How Cassidy automates invoice processing using AI

Step 1: Trigger on invoice intake

The Workflow activates when a new invoice arrives—whether via email inbox, e-invoicing API, or direct upload—capturing the document and associating it with the correct client entity and project context.

Step 2: Extract and normalize invoice data

Cassidy's AI extracts header and line-item details including vendor, dates, PO numbers, amounts, tax, and terms. Vendor-aware learning remembers layouts and flags duplicates or anomalies for review.

Step 3: Auto-code to GL with multi-dimensional tagging

Using your Knowledge Base of chart of accounts, vendor defaults, and coding policies, Cassidy assigns GL accounts plus dimensions like department, class, location, job/project, and grant—handling fixed-asset capitalization and prepaid logic automatically.

Step 4: Route for SoD-safe approvals

Cassidy applies your approval matrix to route invoices based on vendor, amount, entity, and project. Approvers receive contextual notifications with PO details, coding, and attachments—with bill approval and payment authorization kept separate to enforce segregation of duties.

Step 5: Orchestrate payment and sync to ledger

Once approved, Cassidy batches invoices for payment by due date and discount windows, generates payment files (ACH, wire, check, or virtual card), and syncs payment status and references back to your accounting system for instant reconciliation.

Step 6: Log audit trail and surface visibility

Every step—who approved what, when, and why—is captured in an immutable audit log. Dashboards surface cycle time, first-pass yield, exception rates, and on-time approvals across your entire client portfolio.