NRC Health
Huey: one AI presence across a whole platform
Huey is NRC Health's AI assistant. As it spread across products, each team was inventing its own way to surface AI. The goal: unify how users meet and use AI across the whole platform.
- Role
- Lead Product Designer
- Team
- 2 product designers, 1 UX researcher
- Products
- Results dashboard, engagement tracking, rating scorecards, regulatory surveys, feedback management
- Deliverables
- Touchpoint map, 3 principles, evaluation test, marker and drawer
01
My role
Led the research workshop with the design team
Framed the problem along two axes: who triggers Huey, and what the user is trying to do
Synthesized the touchpoints into three design principles and wrote them
Turned the principles into an evaluation test and a first UI pattern
02
Mapping touchpoints
We mapped where users would want Huey, what should trigger it, and in which context: exploring data, monitoring, or problem-solving. The answer to context was "all of them". So the question became how Huey shows up, not where.
Enhanced search
Chat and explore, insights from data
Manual
"Explain this", "Dive into this" on contextual items
Aggregator
Correlate data, create summaries
Automatic
Summaries and alerts inside existing workflows
All of them.
Exploring data, monitoring, problem-solving.
03
Three principles
The touchpoints collapsed into three qualities every Huey interaction must have. Each one got a visual signifier, so the principles could live in the UI and not only in a document.
#1
Inviting
Always reachable, never intrusive
Huey invites engagement through clear, discoverable patterns without demanding attention. An approachable presence that supports rather than interrupts.
#2
Relevant
On point and user-centered
Huey reads data state, workflow stage and task intent. AI appears when it adds value, with actionable insights and next steps rather than generic answers.
#3
Adaptive
Adjusting its level of presence
Huey integrates into existing workflows through familiar UI patterns. It adapts presence and persistence to context, fitting the tool to the user's intent.
04
Principles as a test
The principles verify designs rather than being bent to fit them. Any Huey marker, feature, pattern or interaction has to pass all three to be ready to integrate.
Is it relevant to this specific context and user need?
RelevantIs it inviting without being intrusive or demanding?
InvitingDoes it adapt to existing workflows rather than create friction?
Adaptive05
From principles to UI
The signifiers became the Huey marker, explored in weight and color variants. The drawer gives every product the same anatomy for AI conversations.
Actions
More, pop out, close
Disclaimer
Transparency on AI
Content
Answers, insights, next steps
Feedback
Rate each answer
Search
Ask in natural language
06
Outcome
The principles became the shared test for any Huey marker, feature, pattern or interaction, and the drawer gave every product the same anatomy for AI conversations. Under the same Huey umbrella, ambient listening raised nurse documentation from 55% to 90%.