End-to-end Automation of SAP Purchase Info Record creation and validation using RPA and AI Orchestration
Vu, Trung Dinh (2026)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202604308682
https://urn.fi/URN:NBN:fi:amk-202604308682
Tiivistelmä
This thesis examines the feasibility of automating SAP master data processes, focusing on Purchasing Info Record (PIR) creation and validation. The current process is largely manual, involving data extraction from Salesforce, data entry into SAP, and manual validation, leading to inefficiencies and risk of errors.
The objective is to design and implement an automated solution combining Robotic Process Automation (RPA) with an AI-assisted interface. The system consists of a Copilot-based user interface, an orchestration layer for request routing, and an RPA execution layer handling data retrieval, processing, SAP entry, and validation.
The implemented solution automates data extraction, transformation using Excel, and SAP interaction through scripting. Validation is performed by comparing SAP data with source data, with exceptions escalated to human users.
The results show a significant reduction in manual workload and an estimated 40% decrease in processing time. The solution demonstrates scalability and can be extended to other SAP master data processes, although limitations remain in complex cases and system dependencies.
The objective is to design and implement an automated solution combining Robotic Process Automation (RPA) with an AI-assisted interface. The system consists of a Copilot-based user interface, an orchestration layer for request routing, and an RPA execution layer handling data retrieval, processing, SAP entry, and validation.
The implemented solution automates data extraction, transformation using Excel, and SAP interaction through scripting. Validation is performed by comparing SAP data with source data, with exceptions escalated to human users.
The results show a significant reduction in manual workload and an estimated 40% decrease in processing time. The solution demonstrates scalability and can be extended to other SAP master data processes, although limitations remain in complex cases and system dependencies.