Product & systemsMelodyArc

Designing a multi-role service platform from the ground up

An embargo-safe account of the systems, workflow, and delivery decisions behind more than 250 production screens and flows.

Role
UX Design Lead
Duration
July 2023–present
Team
Product & engineering
Scope
250+ screens & flows

Challenge

MelodyArc is a service-based B2B company that brings data, rules, AI, and people into operational workflows. Its official documentation describes Portal as a web workspace where teams manage conversations, review progress, intervene in workflows, and administer their organization.

The design challenge was to turn that broad service model into a coherent product: many roles, tasks, states, permissions, and operational decisions needed to feel like one understandable workspace rather than a collection of disconnected tools.

Evidence

The most useful public evidence is the scale and structure of the shipped work. I designed the Portal from the ground up across more than 250 production screens and flows, while building reusable patterns that could support continued product growth.

250+
production screens and flows
3
documented Portal personas
1
shared product language

Documented surface area

MelodyArc’s official documentation names three Portal personas—Associate, Designer, and Expert—and describes the following areas. They are listed here to communicate product breadth, not private implementation detail.

Conversations & tasks

Active work, progress, context, and actions within a shared operational frame.

Knowledge & expert tools

Documented spaces for knowledge access and expert contribution.

Monitor

Visibility into work that needs review, awareness, or intervention.

Organization administration

Documented users, roles, operators, views, and organization settings.

Decisions

  1. Keep work and context together. Operational tasks need their conversation, current state, supporting context, and available actions close at hand.
  2. Design for roles and permissions. A multi-role product cannot rely on one universal view; hierarchy and available actions must stay legible as responsibility changes.
  3. Systemize repeated states. Loading, empty, active, blocked, error, review, and completion behavior were treated as reusable product infrastructure.
  4. Make progression traceable. Status, ownership, and the relationship between actions and outcomes need to remain understandable across a long-running workflow.

Operating model

Model the work

Map roles, task progression, decision points, permissions, and required context.

Build the system

Turn recurring hierarchy, component behavior, responsive patterns, and states into a shared foundation.

Ship with engineering

Resolve behavior through implementation and evolve the system as new operational needs appear.

Outcome

The work established a production product across 250+ screens and flows, supported by reusable components, interaction states, responsive layouts, and a shared visual and behavioral language.

This account intentionally makes no claims about confidential customer outcomes or unpublished business metrics.

Reflection

At this scale, design quality depends less on perfect individual screens and more on the coherence between them. The strongest decisions were the ones that helped new workflows inherit familiar structure while still making their specific operational responsibilities clear.

Interested in the platform work?

I can discuss the approach privately while respecting the product embargo and confidentiality boundaries.

Start a conversation ↗