There was no way to preview the outcome of a workflow before it went live. Teams had to save, wait, and hope the alert landed in the right place.

Users couldn't tell how filter logic worked or what a given field actually controlled. This led to trial-and-error configuration — risky in a production alerting system.

Workflow configuration involves multiple independent decisions: filters, destinations, notification logic. Instead of a single long form, we broke this into sequential steps. Each step has one clear job, which reduces cognitive load and makes it harder to skip something important by accident.

The biggest source of uncertainty was the destination step: what will the Slack message actually look like? I added a live preview that updates as users configure their notification settings. This directly addresses the "save and hope" problem — users can verify the output before anything goes live.

Filter logic in alerting tools is genuinely complex — terms like "enrichment" or "muting rules" aren't self-explanatory. I added contextual descriptions directly inside the form, visible without any extra click. This makes the form readable for users who are configuring a workflow for the first time, without slowing down those who already know what they're doing.

In New Relic, each team has a central hub page: a single place that brings together everything relevant to that team. Workflows live there too. This meant the workflow summary wasn't a standalone screen dropped into a global menu, but a natural part of how teams already navigate the product. Users arrive at their workflows in context, alongside the alerts and entities they're already managing.