a compliance dashboard that cut diagnosis time by more than half
Admins had to open each environment to find violations. I led the design of a dashboard that showed which environments were blocked and why, right in the table, with the fix one click away.
4–5 min → 1–2 min
Admins took about 4–5 minutes to find why an environment was blocked. In the new flow, it took 1–2 minutes, with a suggested fix alongside.
Remediation plan in 1 click
The old flow gave admins no remediation plan. Now one click opens the plan, and in the full design, one more sends it to the developer.
6 of 6 admins asked for it
Every admin we tested with wanted the new flow. One asked, “Can we get this, like, yesterday?”
the problem
admins could see their environments, but not what was wrong with them.
We were at risk of losing two major clients. The existing environment list gave admins no status, so they had to open each environment by hand to spot risks and decide what to fix first. That cost their companies time and money, and it frustrated the admins doing the work.
constraints
Limited scope
Limited engineering capacity
Tech debt in the old UI
A metrics-only starting point
decision 01 · reframing the direction
show why an environment is blocked, not how many are.
the constraint
The ask was to build a dashboard, and our PM’s prototype put measurements in one place with no story behind them.
the decision
I proposed a table that ties every blocked environment to the policy rule it broke, so admins see the reason at a glance and can act on it. Our PM was fully on board.

Our PM’s metrics-first prototype: counts with no reason behind them.

My wireframe: the table-first layout at a high level, before I researched the charts.

Before launch · Admins opened individual environments to check whether a policy rule had been violated.

Proposed workflow · Start with the blocked environment in the table, identify the rule it broke, then follow through to a fix.
decision 03 · protecting the design
hold the line on figma as the source of truth.
the constraint
Builds kept drifting from the designs. Design changes landed in the sandbox without engineers alerting me for review.
the decision
I started checking in with engineering twice a week, so builds came to me as soon as they were ready. From then on, builds came straight from Figma.
the tradeoff
I met engineering halfway on the menu bar. Moving it from the old UI to the new one carried real tech debt, so we shipped part of the update and I ticketed the rest for iteration 2.

The Figma design.

The sandbox build before I reset Figma as the source of truth. The red boxes mark where it drifted.
ux research
admins wanted the answer in as few clicks as possible.
The old experience was hard to use, so admins wanted the new one to be quick and simple. I turned what we learned into two concepts, wireframed each, and built a Claude prototype of both so admins could try them for real.
two concepts
two bets on what admins need first.
I developed both concepts into wireframes, exploring how admins could scan compliance, investigate an environment, review alerts, and track changes over time.

Concept 1, “scan the status, explore the details”: summary cards and one environments table, with drawers for investigating. Admins found the table easier to scan, so this direction won.

Concept 2, “focus on what needs attention”: environments grouped by status, with compliance trends at the top.
two rounds of testing
six admins, two rounds.
In round 1, six admins across regions compared both concepts in my Claude prototype. In round 2, we ran moderated usability testing on the preferred design: the ideal flow from creating a policy to fixing a blocked environment.

Round 1: admins comparing the two concepts. They preferred a mix of both.
Round 2 · Usability-tested admin prototype. The walkthrough shows the compliance trend and how admins filter to blocked environments to investigate a violation.
what admins told us
simplify the table, make everything visible
Admins wanted to see everything, including the fixes, at a glance. I condensed the separate violation columns into one, shrank the status badges to icons to free up space, and added a suggested-fix column. Now admins hover over a fix to see why an environment is blocked, and one click opens the drawer.

too many clicks to find a fix
Clicking an environment took admins to a separate page that listed its problems in a table. They wanted the fix right away. I designed it so clicking an environment opens a drawer with the commands to fix each violation.

help us reach our developers
Admins wanted a quick way to message developers in Slack, and one asked for Jira. In the drawer, I added a way to copy all the fix commands at once, or select them and send them to the developer. Sending to Slack and Jira is part of the North Star.

outcome
what shipped, and when.
The first release launched the environments table with block reasons, alerts, and compliance trends. The remediation drawer, with its one-click fix plan, shipped in a later release, and the remaining menu updates were deferred. Early results came from a beta with two companies.
4–5 min → 1–2 min
Time for admins to find why an environment was blocked, down from about 4–5 minutes in the old flow.
~9–10 → 1 click
To get from a blocked environment to the commands that fix it, down from about 9–10. Hovering in the table shows why it’s blocked.
6 of 6 admins asked for it
Every admin we tested with wanted the new flow. One asked, “Can we get this, like, yesterday?”
final design
what’s next
exploring the developer side of remediation.
Admins asked for a quick way to reach developers in Slack. I explored how developers might receive those issues and see what to fix. Developer notifications and the Slack handoff were deferred from the first release.

Future exploration · Early developer view for investigating and resolving environment issues.
reflections
leading meant looking beyond the screen.
This project taught me to own the full experience, from shaping the workflow to protecting it through build.
lesson 01
make QA a habit
Implementation drift showed me that handoff needs regular follow-through. Checking in twice a week caught discrepancies while the work was still in progress.
lesson 02
keep the project moving
I learned to take a more active role in coordinating delivery, starting regular check-ins to surface blockers, clarify next steps, and maintain momentum.
lesson 03
follow the whole journey
I initially focused too closely on the dashboard. Testing pushed me to consider what happened before and after it, including how admins passed issues to developers for remediation.
next case study