Your company is in the process of designing a data lake on Google Cloud Platform (GCP) and aims to develop diverse ingestion pipelines to collect unstructured data from various sources. Once the data is stored in GCP, it will undergo processing through several data pipelines to create a recommendation engine for end users on the company's website. The structure of the data from source systems can change at any time, necessitating a solution that can handle these changes. Additionally, the data must be stored exactly as it was retrieved to allow for reprocessing, should the data structure be incompatible with current processing pipelines. Given these requirements, how should you design the architecture to support this use case after retrieving the data?