Labxiid · Enterprise AI · 2026
Designing core product flows for an enterprise AI startup. Part-time, from scratch, shipped to pilot in 2 months.
My role
Product Designer · Contract
Timeline
Feb - Apr 2026
Scope
0-to-1 · Enterprise AI
Outcome
Live pilot · B2G
Overview
A friend was building an AI startup and needed design help. I took it on part-time — about 4 hours a week over 2 months — designing the core product flows while his engineering team built the product.
Labxiid AI is an enterprise AI workspace for organizations that can't use public cloud AI — banks, hospitals, and government agencies where all data has to stay inside a closed network. The product is now in live pilot with enterprise clients including 농협중앙회 (NongHyup Bank) and 고대의료원 (Korea University Medical Center).
My role
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
My role
Content Studio
AI-generated reports, presentations, and documents using internal data.
My role
Enterprise Data Library
Centralized library of organization's data sources,
What I Designed
All core flows — navigation, search, content, and data management.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Deep Search Agent
Natural language queries against internal databases and documents, with source citations.
Key design decision
Legible hierarchy over maximum flexibility
The core conflict
The engineering team's instinct was to make everything available in every context. On paper that sounds ideal — maximum flexibility, minimum friction. In practice it collapses the experience.
When everything is available everywhere, users lose their sense of place. They can't feel what level they're operating at, or what belongs here versus somewhere else. There's no hierarchy — just a flat, overwhelming surface where everything competes equally.
The specific conflict: resource scoping
The original proposal: changes to resources inside a chat would automatically propagate to the project level. A single source of truth. I pushed back — a chat feels like a contained, in-the-moment workspace. A project feels permanent and shared. Silently connecting the two breaks the user's mental model.
We landed on scoped actions: chat-level changes stay in the chat. Project-level changes require a deliberate action at the project level.
The second principle
Always be clear about what will happen before a user takes an action, and what is happening while they take it. In enterprise software handling sensitive financial and medical data, invisible side effects aren't a UX inconvenience — they're a trust failure.
These two principles — legible hierarchy and transparent actions — were the foundation of every significant design decision on this product.
Impact
Shipped to pilot in 2 months
All core flows designed and handed off part-time. Product went from no design to live enterprise pilot.
Live with regulated clients
농협중앙회 and 고대의료원 — Korean banking and healthcare, two of the most compliance-sensitive industries.
Scoping architecture established
Chat-level and project-level actions deliberately separated — protecting users from invisible side effects in sensitive data environments.
Full product scope
Navigation, search, content generation, data management, and resource flows — all designed and shipped solo.
Key insight
Designing for enterprise AI is mostly about trust.
Users working with sensitive financial and medical data inside a closed network are already skeptical of AI. Every design decision has to answer the same question: do I know what's happening? Can I trust what I'm seeing?
That's why hierarchy and transparency matter beyond the interaction detail. A product that feels predictable — where users understand their scope, know what their actions will do, and never encounter invisible side effects — is a product people will actually use. In a regulated environment, that's not a nice-to-have. It's the whole job.