ORA ClientTalk
Reducing cognitive load in AI Coaching through semantic data visualization
Timeline
January 2025 - May 2025
Company
Mentivista (Client: ORA Space), an AI-powered LMS for beauty & wellness businesses. This project focused on the 'Shadowing Training' feature, enabling managers to review transcripts and coach sales representatives on compliance and empathy.
Role & Team
Founding product designer, collaborated with 1 PM, 1 AI Engineer, and 2 Software Engineers.
Responsibility & Collaboration
First design hire, led the end-to-end redesign and architected the semantic interface for AI-driven risk & empathy auditing.
With PM, defined MVP strategy and scope by prioritizing the visualization of invisible FDA risks and soft-skill gaps to ensure regulatory compliance.
With AI/Software engineers, built a Semantic Design System, mapping latent model outputs into a deterministic visual hierarchy to translate opaque data into actionable heuristics.
Impact
0%
Critical FDA Miss Rate
-85%
Back Tracking Rate
3.5x
Coaching Velocity
User Problem
Managers struggle to distinguish critical FDA risks from minor soft-skill gaps due to flat, unhierarchical data.
This lack of hierarchy forces managers to spend excessive time searching for context, increasing the risk of missing legal liabilities.
As-Is Interface
😰
"It’s nerve-wracking because critical FDA violations look identical to minor soft-skill feedback. When I’m busy, I’m terrified I might miss a legal liability just by skimming through. I end up re-reading the entire transcript anyway just to be safe."
_ Senior Salon Manager
😖
"AI is supposed to help, but with no hierarchy, I still have to cross-check everything myself. This data dump is exhausting to look at. The insight and evidence feel so disconnected that I feel like it's taking me longer than doing it manually. "
_ Operations Lead
Business Goal
Our business goal was to maximize coaching velocity by bridging the validation gap with a 3-layer semantic token system that instantly prioritizes legal risks over soft-skill feedback
Solution
A semantic traffic light system for risk & empathy
01. 3-Layer Semantic Tokens
Distinguishing binary legal risks from nuanced soft-skill gaps using a strict color hierarchy
02. Context Anchoring Module
1:1 evidence connection, physically connecting the AI insight to the raw transcript evidence to eliminate manual searching.
Research - User Journey
While Speech-to-Text and AI evaluation were implemented, the training efficiency didn't improve. Why? The AI dumped data without hierarchy, causing "Cognitive Confusion" for managers.
How might we reduce cognitive load for managers to identify critical risks instantly?
Ideation - Approaches and Goals
Problem
Invisible Risks
FDA violations and critical risks are hidden within plain text, making them easy to miss during manual review.
Approach
Semantic Traffic Light System
A strict 3-tier color system (Red/Yellow/Green) that visualizes the urgency of each insight based on regulation and impact.
Successful Metrics
0% Miss Rate
in Critical FDA Violations
Disconnected Context
Verifying subjective "Soft Skills" (e.g., Empathy) requires searching through the whole transcript, causing high cognitive load.
Context Anchoring Module
An interactive feature where clicking an insight auto-scrolls and highlights the exact evidence in the transcript.
-40% Search Time
Reduced Verification Time
Iteration 1: Semantic Clarity — The Traffic Light System
1) Comparison
Before: High cognitive load with flat text.
After: Instant recognition with semantic tokens.
2) Design Decision
Why Tokens?..
"Signal vs. Noise" Initially, I explored using Full-Background Colors to maximize visibility. However, user testing revealed that this created "Alarm Fatigue"—when everything is colorful, nothing stands out. I pivoted to a Minimalist Token System (Left-border + Icon), which reduced visual noise by 60% while maintaining the same level of urgency recognition.
❌ Rejected
✅ Selected
3) Solution
Visualizing the Invisible Weight of Information
From Flat List to Hierarchical Cards. Previously, a critical FDA violation looked the same as a minor typo. I implemented a Semantic Color Token System to instantly signal urgency without requiring managers to read the details first.
Iteration 2: Efficiency — Context Anchoring
1) Observed Behavior : The Scroll-Hunting Loop
Unlike binary FDA violations, feedback like 'Lack of Empathy' is abstract. We observed managers reading the transcript up and down repeatedly to judge if the AI's assessment was fair. This created a massive validation gap.
👥
"The AI says he was 'lack of empathy', but I can't see the context. I spent more time finding the sentence than actually coaching him."
2) Solution
Bridging the Gap Between Insight and Evidence
One Click, Instant Context. I designed the 'Auto-Scroll & Highlight' interaction. Clicking an evaluation card instantly navigates the transcript to the exact timestamp, visually connecting the Result (AI Insight) with the Cause (Raw Data).
Deep Dive : Accessibility — Designing for Every Manager
A) Double-Coding for Color Blindness
By pairing colors with distinct icons, I created a fail-safe system where information is conveyed through both hue and shape.
B) Invisible Hierarchy for Screen Readers
translated visual importance into Structural Code (Semantic Markup & ARIA Labels), ensuring that screen reader users perceive the same urgency for critical risks as sighted users.
Retrospective & Conclusion
This project highlighted that in AI-native design, Semantic Tokens are the essential infrastructure of user trust. I realized that a 'Flat Data Dump' without a clear hierarchy inevitably traps managers in a 'Scroll-Hunting Loop' as they struggle to verify abstract AI feedback against raw evidence. By engineering a 3-Layer Semantic Token system, I moved beyond mere aesthetics to encode urgency directly into the UI, using a strict logic that instantly separates Critical FDA Risks from secondary behavioral observations. This system was vital for bridging the 'Validation Gap,' transforming the interface into a prioritized decision-support engine that allows managers to focus on high-stakes coaching rather than manual data auditing.