Michael Cohen
← All Work

Replacing a broken manual close with an automated reconciliation engine

Read the full case study →

Detailed write-up · Methods · Full outcome narrative

Tools & Methods

PythonPandasElectronReactTypeScriptMicrosoft Dynamics NAVPower BIExcel

The finance lead was re-importing ERP data up to five times in a three-hour window at month-end. A single person handled all bank postings for every entity. The close took weeks. The root cause was process design — not capacity.

Scoped, architected, and delivered an Electron desktop application with a Python/Pandas backend performing a 3-way match across payment processors, internal platform records, and bank deposits. Built direct ERP connections into close workbooks and connected Power BI for live GL visibility.

3–4 hour manual reconciliation replaced with an automated engine · 4 payment processors covered: PayPal, Stripe, Braintree, NMI · Manual GL export loop eliminated via direct Dynamics NAV connection

Working on something similar?

Every engagement starts with a conversation about where your finance or data function is breaking down. No intake forms, no pitch.

← Back to work