Scaling agile at RELEX : a case study of the REXUS framework
Rehman, Ayesha (2026)
Diplomityö
Rehman, Ayesha
2026
School of Engineering Science, Tietotekniikka
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20260714111042
https://urn.fi/URN:NBN:fi-fe20260714111042
Tiivistelmä
Large software organisations often struggle to scale agile beyond individual teams. Ready-made frameworks like Nexus, SAFe, and LeSS exist to help, but many organisations end up building their own. This thesis examines one such case: REXUS, a custom scaled agile framework built and rolled out at RELEX Solutions, a Finnish supply chain software company with around 800 R&D engineers. REXUS started from Nexus but grew into something quite different, blending elements from multiple frameworks to fit RELEX's specific needs. Its launch in February 2024 coincided with a major organisational restructuring, making the two inseparable in practice. Data was collected through internal document analysis and twenty semi-structured interviews across five role groups.
The study found that REXUS came about because existing ways of working had broken down i.e. planning sessions were too long, dependencies were caught too late, and there was no shared rhythm across teams. The framework brought real improvements: better visibility, more reliable planning, and a shared cadence that the organisation had lacked. But dependency management, one of the main reasons REXUS was created, remained a stubborn problem because it was rooted in how the organisation was structured rather than how it planned. The framework was also applied more rigidly than intended, and adoption varied considerably across teams. The thesis closes with an unexpected turn: during data collection, RELEX announced that REXUS would be discontinued as part of a company-wide restructuring driven by the rise of AI in software development. The case ultimately documents both what a bespoke agile framework can deliver and what is left unresolved when one is dissolved rather than evolved.
The study found that REXUS came about because existing ways of working had broken down i.e. planning sessions were too long, dependencies were caught too late, and there was no shared rhythm across teams. The framework brought real improvements: better visibility, more reliable planning, and a shared cadence that the organisation had lacked. But dependency management, one of the main reasons REXUS was created, remained a stubborn problem because it was rooted in how the organisation was structured rather than how it planned. The framework was also applied more rigidly than intended, and adoption varied considerably across teams. The thesis closes with an unexpected turn: during data collection, RELEX announced that REXUS would be discontinued as part of a company-wide restructuring driven by the rise of AI in software development. The case ultimately documents both what a bespoke agile framework can deliver and what is left unresolved when one is dissolved rather than evolved.
