According to some commentary, employees are empowered by simple, flexible and lightweight social software such as blogs and wikis. Using such software, it is claimed they will be able to connect and collaborate beyond boundaries and to benefit organisations in innovation and growth. Extending such claims, commentators also argue that social software challenges existing ways of networking, communicating and collaborating and is likely to cause disruption to some organisations. However, a literature search reveals that it is still unclear for both researchers and practitioners how social software can be exploited for work purposes, and what are the grounds for its adoption. Systematic investigation of the adoption of social software for work collaboration is found to be absent in academic writing.Organisations continue experimenting on social software, seeking ways to encourage their employees to adopt the tools, and expect to benefit from self-organised, distributed collaboration. This technology adoption is a contingent process. The way members use social software is embedded within the social interaction process, group characteristics, tasks and environments.Adaptive structuration theory (AST) is adopted for social software-supported collaboration in a way that emphasised its value as a contextual analysis approach. A conceptual framework was developed from the researcher's experiential knowledge, AST and empirical evidence, and then revised with the findings from exploratory studies.The main theoretical insights induced from empirical evidence are: influence from "meta-users", e.g., supervisors and managers, affects both users' use intention and behaviour. Social structures embedded in social capital can enable or inhibit the adoption of social software. Nevertheless, new social structures do emerge from online conversations. Additionally, the results show that task-technology fit has little relevance to social software adoption.A case study approach with mixed methods is adopted in this study. Social network analysis and statistical analysis provide complementary support to qualitative analysis. The UK public sector was chosen as the research context. Individuals are knowledge workers in distributed and cross-boundary groups. The asynchronous social software applications studied are blogs and wikis.