Prompt-Based REST API Test Amplification in Industry: An Experience Report

TL;DR

This paper reports an industrial replication of LLM-based REST API test amplification within a large Belgian logistics company. By applying the same amplification approach used in prior open-source work to a production microservice and six endpoints, the study demonstrates meaningful gains in structural API coverage while navigating industry constraints such as authentication, statefulness, and tooling policies. It finds that amplification is helpful but requires careful integration with existing test practices, explicit state management, and human oversight, with manageable post-processing effort. The work contributes practical lessons for practitioners and highlights challenges—specification scoping, tool adaptation, and realistic automation limits—when translating open research into industry practice. Overall, LLM-based amplification can strengthen industrial API testing when applied selectively to rich endpoints and integrated into existing workflows.

Abstract

Large Language Models (LLMs) are increasingly used to support software testing tasks, yet there is little evidence of their effectiveness for REST API testing in industrial settings. To address this gap, we replicate our earlier work on LLM-based REST API test amplification within an industrial context at one of the largest logistics companies in Belgium. We apply LLM-based test amplification to six representative endpoints of a production microservice embedded in a large-scale, security-sensitive system, where there is in-depth complexity in authentication, stateful behavior, and organizational constraints. Our experience shows that LLM-based test amplification remains practically useful in industry by increasing coverage and revealing various observations and anomalies.