01
The problem
A large global travel technology company running NetSuite, Salesforce, Expensify, and a custom middleware layer nicknamed "Pontus" for intercompany reporting had four compounding problems: a month-end workbook process of thousands of clicks through NetSuite saved searches, Excel manipulation, and intercompany reconciliation done the same manual way every cycle; a middleware layer with a single point of failure, if the one engineer who built it left, the system would be stuck; Salesforce-to-NetSuite billing and rev rec described internally as "super manual" despite the company's scale; and Expensify reporting split across multiple files per entity and currency requiring significant manual updates.
02
What Loopfour automated
- Month-end workbook automation: NetSuite saved searches flow into data manipulation and reconciliation, producing a formatted workbook ready for review and approval, the same process every month, with no Excel downloads and no manual manipulation
- Intercompany and reporting (replacing the middleware layer): revenue data flows into allocation by origin and out to intercompany entries, with no dependency on one engineer and a full audit trail
- Salesforce billing automation: Salesforce opportunities flow into invoice generation and a rev rec schedule, automating the billing side of Salesforce without ripping it out
The build philosophy here is "understand, don't mimic": the system doesn't record clicks, it understands what the workbook is doing and replicates the logic, with every step showing its reasoning so the controller can trace from source data to output. Nothing posts without review.
03
Why it matters for controllers
A close process that depends on one engineer's tribal knowledge is a close process with a single point of failure sitting in your control environment. This one replaces tribal knowledge with an inspectable, repeatable rule.
Running month-end close through a fragile, single-owner middleware layer?
Book a demo