A Metaprotocol For a Family of Rapid Multivocal Reviews of Generative AI in the Software Industry

Context: With the growing interest in solutions based on Generative Artificial Intelligence (GenAI) applied to SE, a need to identify solutions beyond conceptual proposals and prototypes arises. In this rapidly evolving landscape, exploring methodological strategies that enable the identification of GenAI-based solutions used in the industry becomes essential. Goals: In this paper, we report on the experience of developing a metaprotocol for a family of Rapid Multivocal Reviews (RMRs) to identify operational GenAIbased technologies being used in the software industry (i) across the phases of the software development lifecycle and (ii) in the planning and executing SE empirical studies. Method: A metaprotocol was developed through an interactive process combining GenAI’s creative support with software engineers’ feedback. GenAI assisted in defining the stages of the RMRs family, while discussion cycles with software engineers contributed to refining and validating it. Results: Six RMR instances have been organized by covering the perspectives of Software Requirements, Design, Test, Coding, Management, and Empirical Methods described in SWEBOK 4.0. It allowed guiding the execution of the RMRs and highlighted the challenges of organizing the metaprotocol and justifying the decision-making. Conclusion: Using a metaprotocol to support the observation of industrial experience using GenAI in software projects through RMRs will contribute to organizing a body of knowledge regarding GenAI-based solutions currently available and used in the software industry to support the SWEBOK 4.0 practices.