AQCLab is a laboratory for software and data quality evaluation in conformance to the ISO/IEC 25000 series of standards. As such, requirements are a fundamental element in the evaluations that are carried out, as they are the input for software Functional Suitability and Data Quality, two of the types of evaluation carried out by the laboratory. Software Maintainability evaluations are also carried out by the laboratory, where the requirements on applicable metrics and thresholds to be met have been established by the laboratory.Furthermore, the laboratory's personnel possess considerable expertise in the domain of auditing software lifecycle processes, which has enabled them to observe how requirements are managed in multiple companies.In this article, we present the experience of AQCLab over the years, in the capacity of software product and data quality evaluators and software development process auditors, with regard to the manner in which our clients manage and define requirements. This has led us to the realization that these practices are often neglected.

In an era dominated by technological integration, Artificial Intelligence (AI) is pivotal across various sectors, driving significant advancements and demanding robust quality measures for its implementations. This paper introduces a novel AI maturity assessment framework designed in alignment with International Standards provided by ISO (International Organization for Standardization) and IEC (International Electrotechnical Commission), specifically with the ISO/IEC 33000 family of standards for process assessment. Our framework aims to provide AI system developers with a structured tool for continuous improvement of their development processes, thereby enhancing the reliability and efficacy of AI applications. To demonstrate the applicability of our proposed framework, we have validated it through a case study in the automotive sector. Specifically, the framework was employed to assess and enhance an AI project to develop a mechanism to determine vehicle behavior from sensor data within the constraints of onboard devices. Our findings identify key improvement points contributing to the iterative enhancement of AI system quality in engineering applications.