HumanFirst - NLU Disambiguation Redesign
Evolved a mission-critical enterprise workflow without breaking the patterns power users relied on.
Overview
HumanFirst needed to evolve its NLU platform toward generative capabilities while keeping enterprise clients productive in their existing disambiguation workflows. The goal was to add clarity and speed without forcing teams to relearn the core labeling patterns they depended on.
My Role
I led the UX/product redesign, mapping the critical workflow, defining a cleaner information hierarchy, and pairing with engineering to ensure the new model fit existing enterprise constraints.
The Challenge
The platform had to modernize its disambiguation experience for a new generation of AI tooling without disrupting enterprise workflows that were already deeply embedded in production.
Key constraint: maintain familiar interaction patterns while enhancing clarity and decision support.
The Solution
I transformed a cluttered interface into a purpose-built disambiguation workspace that clarifies context, improves hierarchy, and accelerates conflict resolution.
- Clear indication of disambiguation mode
- Visual distinction between conflicting intents and utterances
- Redesigned conflict percentage visualization
- Focused layout for faster resolution
Before & After
Impact & Results
Quantitative outcomes
- Zero workflow disruption during rollout
- 100% client migration success
- Preserved enterprise relationships
Qualitative outcomes
- Enabled major client platform adoption
- Simplified decision-making for disambiguation
- Improved user confidence in conflict resolution
Built with feedback. Not ego.