The end of agile as we know it

Artikel i vetenskaplig tidskrift, 2027

Context: Ever since the Agile Manifesto was introduced in 2001, the agile principles have supported software development teams in continuous delivery of valuable software. With its focus on customer collaboration, and with a promise to help organizations respond to change, it is fair to say that the agile paradigm changed software development and that it, over the years, has become the method of choice worldwide. Objectives: However, despite more than two decades of efforts being spent on adopting and scaling agile practices, companies still do not experience the benefits they had expected. In this paper, we study the challenges companies experience and why existing agile practices fall short in supporting development of large-scale software-intensive embedded systems. Methods: We present a reflective synthesis across a decade of collaborative research within a large research consortium. We do so by synthesizing findings across a selected set of our own prior studies into higher-order interpretations of the challenges that software-intensive systems companies face when implementing, adopting and scaling agile practices. In addition to reflecting on our previous research, we report on a workshop series conducted in 2024 in which we refined and extended the challenges we identified and where we explored what a new paradigm could potentially look like. Results: First, we identify the key challenges that companies in the software-intensive embedded systems domain experience in relation to agile practices. Second, we empirically derive the key principles that a novel, next generation product development paradigm needs to satisfy. Third, we present a conceptual framework, i.e., RADICAL, a strawman proposal for what a such new paradigm could potentially look like. Conclusion: We conclude that for the case companies involved in our research, there is a mismatch between the original agile principles and the reality of product development they face. Hence, there is a need for a new paradigm that supports not only development, but also scaling and monetization of software-intensive embedded systems.