This is where event sourcing comes in, an alternative concept for data persistence. Instead of always saving the current state and regularly overwriting it with updates, event sourcing records the events that led to this state over time – in the language of the domain.

In the fairy tale, these would be events such as: “Princess entered forest,” “Wolf met princess,” “Wolf ate princess,” “Hunter killed wolf,” “Princess was freed,” and so on. Each event is an immutable fact and remains permanently preserved. The current state (such as the princess still being alive or alive again, or her parents missing her) is a projection based on the sequence of these facts.

This approach offers tangible benefits:

• Transparency and traceability: Every change can be traced back to a specific date and cause. You can check at any point in time why a system is the way it is today.

• Clear semantics: Events have technical significance. Instead of generic updates, you see contract renewals, cancellations, role changes, incoming payments – exactly the terms that specialist departments think and speak in.

• Corrections without falsifying history: Errors are not made “invisible” by retroactive manipulation, but are corrected by compensating events. This ensures the integrity of the data line.

• Better analyses: Event streams map behavior over time. This is the basis for process analysis, anomaly detection, and predictions – and thus a solid foundation for applications in the field of artificial intelligence.

• Decoupling and evolution: Events are excellent integration points. Other systems can dock without disrupting the core – a central component of modern, networked architectures.

For more details on how to design and develop software using this approach, visit cqrs.com. If you are particularly interested in the interaction with artificial intelligence – from event recording and projections to features for machine learning models – then we recommend reading eventsourcing.ai.