While cloud computing is not a new technology it is a new way of providing on demand services that is continually and rapidly evolving into different business models. Cloud computing is, however, risky. It involves the processing, transferring and storing of data and databases across different jurisdictions via different cloud service providers, leaving a sense of uncertainty during the transactions. Cloud users may be unaware of the complex technical and legal issues involved. Therefore, providing a framework of trust is the responsibility of cloud providers. However, the concept of trust has many facets and its nuances are beyond the scope of what can be captured in Service Level Agreements (SLA). The main idea I would like to convey in this contribution is to provide a theoretical framework – that I call ‘Plan-like Architectures’ – that sits in between two main theories. On the one hand, the theory submitted by Scott Shapiro in his Book Legality where he adopts Michael Bratman’s account of human psychology that we are all ‘planning agents’. On the other, Lawrence Lessig’s approach which sees the architecture of the Internet as a ‘constraint’ and therefore as a main regulator. The contribution suggests viewing through the lenses of the different interwoven multifaceted attitudes of trust that we encounter in the cloud computing market and argues that this should be streamlined into the SLA architectural framework. This will fill in the gap and remove some of the inherent risks and uncertainties associated with cloud computing transactions.