Agile software development has become widespread, particularly in small team settings, because of its emphasis on user value and adaptability. Scaling agile practices in large organizations, however, presents challenges. As organizations grow, projects become complex, requiring control mechanisms to ensure quality deliverables, align outputs across teams, and connect them to organizational goals. Scaling agile frameworks provide structured approaches to coordinate teams and maintain alignment while preserving team autonomy. Yet, applying these frameworks rigidly can reduce flexibility, overlook organizational context, and disrupt the balance between autonomy and alignment. This study examines how different control mechanisms, both formal controls such as structured processes, rules, and performance metrics, and informal controls such as trust, social norms, and mutual adjustment, can foster organizational alignment while preserving team autonomy in large organizations that implement scaling agile frameworks. A multi-method approach was adopted, combining a systematic literature review with an interpretive case study. The literature review synthesized prior research and produced eight theoretical propositions on the dynamic interaction of control, alignment, and autonomy. The case study was conducted in the IT department of a multinational organization transitioning from waterfall to agile, offering an empirical setting to explore how these propositions unfold in practice. Our findings show that formal controls, such as adherence to architectural standards, provide structure and support alignment. Informal mechanisms, including communities of practice, peer coaching, and visibility-based incentives, enhance accountability and autonomy. However, rigid separation of formal and informal controls is restrictive. More integrated hybrid approaches that combine agile methods with structured processes, supported by parallel decision-making bodies, address inter-team dependencies while maintaining flexibility. Large organizations should adopt context-sensitive strategies when applying scaling agile frameworks. Control mechanisms need to be tailored to organizational needs, maturity, and conditions, supported by clear communication, cultural commitment, and adaptive governance.

Behavior Driven Development (BDD) offers a way to express scenarios, written in structured natural language, on how the system should act to fulfill a requirement. Numerous templates have been written to construct BDD scenarios and various practices were born out of usage. Nevertheless, we mostly fail to find documentation on the used templates and understand empirical practices. This paper explores empirically the BDD templates used in practice by evaluating the keywords mostly associated to the GIVEN, WHEN and THEN scenario dimensions. It then studies whether we can use these existing keywords to build a set of non-redundant concepts covering the representational needs and being usable as reference when constructing BDD scenarios. We consolidate these findings in a meta-model.