One alerting system, three products
Fewer missed alerts, one place to configure them
The problem
Three product teams had each built their own way of notifying users. Every team owned its own bugs, and users who touched more than one product got an inconsistent, sometimes duplicated, experience.
The approach
Rather than pick a winner among the three existing systems, the team built a thin shared layer that every product could call, with delivery, retries, and preferences handled in one place. Each team kept its own triggers and content, but stopped maintaining its own delivery pipeline.
The reusable idea
When three teams have each solved the same problem slightly differently, the fix is rarely to pick the best one. It is usually to find the thin shared layer underneath all three and let everyone keep their own judgment on top of it.