Using big data with serverless architecture for customer journeys presents several unique challenges that impact efficiency and reliability. A primary concern is data integration and processing complexity, as individual ephemeral functions struggle with the intricate ETL and transformation logic required for vast datasets, often leading to resource inefficiencies. Another significant hurdle involves state management across stateless functions, necessitating external databases or caching mechanisms to maintain a persistent view of customer journey context, which adds architectural complexity and potential latency. Furthermore, cost predictability and optimization become difficult; while serverless is pay-per-execution, high invocation counts and large data transfer volumes inherent to big data can quickly inflate expenses. Achieving genuine real-time insights and low-latency processing is also challenging due to potential cold starts and the distributed nature of serverless components handling massive data streams. Lastly, observability and debugging across numerous ephemeral functions processing big data streams are significantly complicated, making it harder to trace issues and ensure data integrity throughout the customer journey.