Integrating AI Tools Into the Continuous Testing Process.

Аннотация

This paper utilizes AI tools to enhance the SUBMITED 18 April 2025 ongoing test cycle in a DevOps environment, thereby ACCEPTED 24 May 2025 PUBLISHED 30 June 2025 creating Metabase. This data architecture is robust and VOLUME Vol.07 Issue 06 2025 scalable, supporting a highly responsive release process. The project is vital since the releases have become more CITATION Vladyslav Korol. (2025). Integrating AI Tools Into the Continuous Testing frequent; standard automation has already reached its Process. The American Journal of Engineering and Technology, 7(06), 222– limit, increasing the costs of maintaining scripts and, 229. https://doi.org/10.37547/tajet/Volume07Issue06-23 consequently, resulting in a significantly higher total cost COPYRIGHT © 2025 Original content from this work may be used under the terms due to the discovery of many more defects. The novelty of the creative commons attributes 4.0 License. of this work is grounded in an approach to choosing and applying AI tools through comparative analysis over available commercial and open-source platforms, supported by content analysis of empirical use cases and quantitative assessments from industry reports. Herein, a methodology is presented that consists of architectural solutions for data lake organization, continuous model training scenarios, and ML endpoints integrated into the CI/CD pipeline, which hosts Predictive Test Selection, as well as self-healing and test- case prioritization mechanisms. It narrates the creation of the prompt-engineer position and the connections between QA/ML experts and organizational facets. This paper discusses the application of clear AI measures for risk assessments. The final calls shown here indicate that intelligent automation enables reducing a regression set to a barely necessary size while maintaining 99.9% bug detection and minimizing false alert failures. This, in turn, leads to improvements in MTTR and TCO quality. TestPilot and FlakeFlagger verify Meta’s practices; furthermore, it is anticipated that forecasts will retain a CAGR of 20.9% in the global AI testing market growth. Solution maturity, encompassing both SaaS and on- premises models, offers a flexible choice to regulated industries. Metabase architecture is shown in which raw The American Journal of Engineering and Technology 222 https://www.theamericanjournals.com/index.php/tajet and processed data are kept separately to ensure timely conducted through a review of 19 works, comprising model retraining as well as to minimize computational three academic papers [4–6], market reports [1, 3, 19], costs. This article will be helpful for software architects, and technical overviews [13-16]. Studies on Predictive QA managers, DevOps teams, and ML engineering Test Selection [4] and LLM-driven test generation [5] specialists involved in building scalable and resilient form the theoretical base. The research on self-healing testing architectures. approaches to reducing flaky failure complements the framework [6]. Industry reports have shown that Keywords: artificial intelligence, continuous testing, traditional automation loses effectiveness as release DevOps, self-healing, Predictive Test Selection, data frequency increases [1, 3] and that the AI-testing market lake, ML integration, software quality, Metabase. is rapidly developing [19]. Additional data were provided on the functional capabilities of market leaders—