Orientation
Users needed to know the state of the migration before choosing an action.
01 / Selected workCyclotron · Enterprise product · 2024
Beam
Turning migration infrastructure into a product operators could actually understand.
Beam automated Okta → Microsoft Entra migrations. My role was to make the system legible enough that operators could understand where they were, what had happened, and what needed attention next.
BEAMHomeDiscoverConfigureDeployCH
MIGRATION / CONTOSO
Good afternoon.
MIGRATION STATEReady to deploy
84% configured
NEEDS ATTENTION12 exceptionsReview now →
IDENTITIES MAPPED2,418
RECENT ACTIVITY
Finance groups imported
3 mappings updated
Deployment check passed
The problem
The automation worked. The mental model didn’t.
The early experience reflected the technical system more than the operator’s workflow. Users had to reason through migration entities, configuration scopes, environments, deployment states, and errors without enough orientation.
The product needed to expose complexity without making users carry all of it cognitively.
Discovery↓Configuration↓Deployment↓Error?↓Figure it out
What changed my mind
The question that reorganized the product
Why are we asking users to understand our data model?
Scale
Navigation that worked for one environment would become painful as tenants and entities multiplied.
Priority
Charts looked impressive. Operators needed status, exceptions, and actions first.
Three decisions
01
Give users somewhere to land
Home became a command center for discovery status, imports, remediation, recent activity, and the next useful action.
02
Make navigation scale before the product does
Grouped global search and clearer environment switching reduced the number of places operators had to remember where work lived.
03
Put action before spectacle
Status, exceptions, and actions moved ahead of trend charts. Operational software has different priorities than executive reporting.
The product model
Before
→- Technical structure
- User interprets system
- Scattered configuration
- Late, contextless errors
After
- Home Migration state + next actions
- Search Entities across systems
- Configuration Operator mental model
- Deployment Progress + contextual recovery
What became real
Design became something engineering could argue with.
Figma gave UX, product, and engineering a shared object for testing hierarchy, navigation, language, configuration, and migration state before those arguments hardened into code.
The immediate output was a stronger conceptual dashboard and product narrative for stakeholder alignment. Later implementation details are only credited where their lineage is traceable.
WireframesHigh-fidelity product modelDashboard narrativeImplementation dialogue
Reflection
What I’d protect earlier next time
Hierarchy problems get more expensive as technical complexity compounds.
A structure that works at small scale can quietly become debt.
What users need to act on should outrank what looks impressive.
Complexity is often necessary. Confusion usually isn’t.