Architecture is the process of solution design. The problem domain needs to be understood, that includes the landscape of the problem and the problem itself.
Then from that understood problem domain, we can create;
- Functional requirements: what the automation needs to do.
- Technical requirements: what kind of performance requirements are there? Data storage redundancy? Geographical regions to cover? And many more questions.
The landscape of the problem domain must also be mapped and understood:
- Internal / External dependencies: what information, and systems need to be integrated to make this solution design work? Think about payment systems or internal systems such as CRM customer data that is required to for such a payment system.
- Policies and regulatory requirements: Internal company policies may enforce data location regulations, such as on-premise requirement of financial data that cannot be in other geopgraphical regions. Or, GDPR as a regulatory requirement means that we need to be able to "scrub" user data from our system while keeping the rest of the data in tact.
Architecture encompasses regulatory, landscape, and system solution design in which Cloud services can play a large role to help implement these requirements significantly faster.