A company stores raw collected data in an Amazon S3 bucket for various analytics workloads. Access patterns are unpredictable and cannot be controlled. The company wants to reduce S3 costs. Which solution meets these requirements?

S3 Intelligent-Tiering is the correct solution because it automatically moves objects between two access tiers (frequent and infrequent) based on changing access patterns, without performance impact or operational overhead. This is ideal when access patterns are unpredictable, directly addressing the requirement to reduce costs while maintaining accessibility. S3 Lifecycle rules are used to initiate this transition. S3 Standard-IA is suitable for data that is accessed less frequently but requires rapid access when needed. However, since access patterns are unpredictable, manually or semi-automatically transitioning to S3 Standard-IA could lead to higher costs if data is frequently accessed after being moved. S3 replication is for copying objects between buckets, not for transitioning storage classes within the same bucket to optimize costs based on access patterns. S3 Inventory helps analyze object storage, but it doesn't automatically transition objects. You would still need S3 Lifecycle rules or a custom solution to act on the inventory data.