Abstract
Agile, lean processes have become the de-facto way to operate in the domain of software intensive products. Methodologies such as the lean startup are reshaping the way new companies and even well-established enterprises seek new opportunities in their operations. In contrast, in research, little has changed during that time - organizations that fund research still expect a solid, linear research plan. In this paper, we present an attempt to challenge this model in software research, based on 7 years of experiences in two large, national, industry-led projects that followed a more agile mindset. Furthermore, we also provide an insight to key learnings and best practices of running software research in agile fashion.