A global payroll platform needed payslips and invoices read correctly across many languages and layouts, because people's salaries depended on them.
This company runs global payroll and employer-of-record services, and its customers have employees working all over the world. That means payslips and invoices from a lot of places, in different languages and in layouts that don't match.
Checking payslips was manual. On the invoice side, the company had built an automation in-house, but it needed constant upkeep.
And these weren't documents where close enough was good enough.
“We deal with complex documents that require a high-degree of accuracy. If salaries do not get paid in the right amount, that's a problem.”
We built extraction that read each payslip and invoice field by what it meant, not where it sat on the page. So there was no template to build for each country or layout, and a new format didn't mean starting over.
Payslips in German, Dutch, Portuguese, Spanish and English all went through the same setup.
The company saw a first demo in June 2023 and signed in September. We were in production by late 2023, and the work ran into 2026.
From January 2024 to August 2026, we processed more than 25,000 pages of payslips and invoices for them.
“Super.AI enables us to deliver our core functionality to customers.”
Director of Operations, global payroll platform
If your documents come from different countries, in different languages and layouts, and keeping templates current has turned into its own job, this is how we handled it for them. Here's more on data extraction from documents.
