Manual data extraction from SAP into Excel for regulatory checks is time-consuming, error-prone, and creates operational risks.
The first step in automation is to thoroughly understand the existing manual clicks, filters, and undocumented preprocessing scripts, using AI tools as an aid.
A new architecture using Python, Pandas, and SAP's REST API was chosen over UI automation tools for its stability and maintainability.
Key challenges included stale test data requiring production comparisons and discovering that the UI and REST API use different, unsynchronized data views.
A parallel run of both manual and automated systems, combined with thorough documentation and CI/CD practices, ensures a smooth and trusted transition.
Automating manual processes eliminates unnecessary steps, and success depends on using stable APIs, understanding system complexities, and building trust with business stakeholders.
Jobs that call for the skills explored in this talk.