The biggest challenge in payroll integration is, without a doubt, that it has to work absolutely reliably and run in bulk within a very short time window. The process also handles highly sensitive data, so security is paramount.

Pamica itself also comes with several specific challenges. Our client runs it on a Windows server that is only accessible under very strict conditions and only within the client's domain. What would scare even a seasoned integrator the most is that the system has no API – no interface that third-party connectors could plug into. Bulk data input is only possible through an import that has to be triggered directly on the server. For the import, it isn't enough to supply the server with a data file; it also needs a specific configuration that sets the parameters of each import and maps the particular data set.

The client processes payroll for many companies, each with its own data source – either a different HR system (for example one that isn't localized for the Czech Republic and is therefore unsuitable for Czech payroll) or they keep data only in files and send it for processing via FTP.

• The target application (Pamica) runs on-premise and has no API
• For security reasons, the integration platform's on-premise agent cannot run on the same server (physical or virtual) as Pamica, so the integration also has to make the two servers talk to each other.
• The import operation on the Pamica server can only be triggered externally by a root user (specific permissions)
• Sensitive data
• A short time window for the whole process, leaving no room for a bulk re-import (which can take several hours in Pamica)
• Until then, communication and hand-off of payroll data was purely manual, between the client's employees and their own clients. During payroll processing, the workload made ad-hoc changes practically impossible.

Of all the integration platforms we know, only Boomi can execute server commands. The choice was clear.

To build a universal solution that would provide Pamica connectivity for third-party applications while being secure and easy to use, we split the integration into two larger parts. The first, more product-oriented, fills in what Pamica is missing – an API. The second connects third-party systems to this add-on API, which anyone can adapt to their own source system and process; it is a completely standard API integration.

The whole process works by turning input data into three files, storing them in specific directories on the Pamica server and then launching them automatically with a server command:

• data files in the format required by Pamica, containing employee data, employment details, pay and holiday or other absences taken
• a configuration file that, once the import starts, sets the right operation parameters and selects the specific data file from storage to be loaded into Pamica
• a launch file that automatically kicks off the entire import

Because the client also needed to enter ad-hoc changes (for example when an employee changes their bank account number at the last minute while payroll is already being processed), we connected the whole process to a change form built with MS PowerApps. With it, a user can manually trigger the integration for a specific employee and push the change into Pamica themselves within seconds.