01 — PROJECT OVERVIEW
Newo AI
Designing the MVP for an Voice AI agent platform
I translated complex AI configuration concepts into clear, implementation-ready product workflows. Working inside a fast-moving product timeline, I shaped the core MVP experience for configuring agents, managing credentials, and navigating a scalable administration platform.
ROLE
Product designer · contract
PLATFORM
Responsive web platform
SCOPE
Agent flows, API keys, patterns, handoff
02 — CONTEXT
An AI product still finding its shape
Newo was developing an early platform that allowed users to create and configure AI-powered agents. The challenge was to organize technically complex functionality—agents, skills, actors, state fields and API credentials—into workflows users could understand, configure and manage through a web interface. I joined as a contract Product Designer to define and deliver the core MVP experience under a fast-moving product timeline.
03 — CHALLENGE
The design challenge
01
Make complex concepts understandable
Users needed to work with unfamiliar entities such as agents, actors, skills and state fields.
02
Support configuration without overwhelming users
Multiple settings, dependencies and system states needed a clear, predictable hierarchy.
03
Move quickly toward an MVP
Evolving requirements demanded rapid prototyping, frequent iteration and close engineering collaboration.
04 — CONTRIBUTION
From requirements to implementation-ready flows
MY ROLE
I turned evolving requirements into structured workflows, reusable interface patterns, and clear interaction details for developer handoff.
My contribution covered information architecture, agent configuration, credential management, form and table patterns, system states, rapid prototyping, validation, and ongoing collaboration with product and engineering.
01 · Information architecture
Defined the hierarchy for key dashboard areas and product entities.
02 · Core workflows
Designed agent skills, API credential management, and supporting states.
03 · Reusable patterns
Created shared forms, tables, cards, navigation, and feedback behaviors.
04 · Handoff
Prepared interaction details and annotated designs for implementation.
05 — PRODUCT STRUCTURE
A scalable model for workspace and agent configuration
The dashboard needed to support account-level administration and the configuration of individual AI agents. I organized the experience around distinct product entities while maintaining a consistent navigation model across the platform.
Workspace
Actors
Agents
Billing
Configuration model
Workspace → Profile and API keys → Actors → Agent → Skills, events, states, settings → Billing
06 — СASE FOCUS: AGENT CONFIGURATION
Guided agent configuration
Agent creation required users to configure several interconnected elements. I structured the experience into focused sections, allowing users to work on one area at a time while keeping the broader configuration visible.
01 · Tabbed structure
Groups related agent settings into predictable, focused areas.
02 · Focused editing area
A central zone for editing the selected skill — one element at a time.
03 · Persistent overview
The skills panel keeps the agent’s structure visible while working.
04 · Clear system actions
Save and Publish carry different levels of visual priority.
05 · Inline guidance
Rules and supporting copy appear next to the fields they describe.
07 — FOCUS: API KEYS
Designing for secure credential management
The API key experience had to balance accessibility with security. Users could generate credentials, identify existing keys, control visibility and status, understand when each key had been created or last used, and manage key states without exposing sensitive information.
01 · Masked credentials
Secret keys are hidden by default after generation.
02 · Visibility control
A key is revealed only when the user explicitly chooses to.
03 · Status toggle
Keys can be deactivated without being deleted.
04 · Usage metadata
Created and last-used dates support auditing and cleanup.
05 · Guidance up front
A persistent note explains how keys are stored and protected.
06 · Single primary action
One clear entry point for generating a new secret key.
08 — UI PATTERNS
Creating consistency across the MVP
Because the product was evolving quickly, reusable patterns were important for maintaining consistency and accelerating the design-to-development process. I applied a shared visual and interaction language across navigation, forms, configuration panels, tables and system states.
09 — FLOW
A simplified agent-creation path
The core flow reduces a technically dense setup into a sequence users can follow: create or open an agent, configure skills, connect actors and state, review additional settings, save, and publish.
Progressive sequence
Each step focuses attention while keeping the overall model understandable.
Clear completion
Save and publish actions make progress and readiness explicit.
10 — OUTCOME
What the work delivered
The project produced implementation-ready MVP workflows for configuring AI agents and managing platform credentials. It also established reusable interaction and interface patterns that could scale as the product expanded.
Implementation-ready
Core agent and credential workflows were ready for engineering handoff.
Reusable foundation
Shared patterns support future functionality without redesigning every state.
Closer collaboration
Detailed interaction decisions reduced ambiguity during implementation.
Scalable structure
The model can grow with additional agent capabilities and administration needs.
11 — REFLECTION
Lessons from designing inside uncertainty
This project reinforced the value of clarifying technical concepts before visual conventions are fully established. The most important design challenge was not visual complexity, but creating a clear mental model that helped users understand how agents, skills, actors, state, and credentials relate.