Executive Summary
Clinical research showed that auditory training is significantly more effective when delivered in a familiar voice, such as a spouse or close family member. However, manually recording personalized content was operationally unscalable.
I led the design of a secure, AI-powered Invite & Clone system that enabled patients to generate personalized therapeutic training using cloned voices from loved ones, and it became a primary differentiator for Lace Pro.
1,000+
Clinics with access to personalized AI voice training
HIPAA + GDPR
Compliant, designed from the ground up for regulated healthcare
0
Manual recordings required: fully automated voice synthesis
The Problem
Training effectiveness improves with familiar voices, but implementing at scale faced critical barriers:
- Manual recording workflows were inefficient and inconsistent
- Clinics could not manage individualized audio libraries at scale
- Scaling across 1,000+ clinics required full automation
- AI voice cloning risked distrust (uncanny valley effect in healthcare)
We needed scalable personalization without sacrificing trust.
The Strategy
Core insight: Emotional familiarity increases engagement and cognitive processing.
Design objective: Build a frictionless, ethically transparent AI voice cloning workflow that:
- Feels safe
- Feels simple
- Feels clinically legitimate
- Scales automatically
Three pillars:
- Seamless invitation flow
- High-quality recording capture
- Transparent AI voice synthesis onboarding
Lace has given me back my youth and joy of enjoying lectures, family, friends and group meetings. I feel so inspired and grateful to understand the process of learning to hear with the brain.
Key Design Decisions
Invite & clone workflow
Patient sends a secure invite to a family member. Clear explanation of purpose and usage. Guided script collection. Removes clinic intervention entirely.
Mobile-first recording UI
Large prompts, environmental guidance (“Record in a quiet room”), real-time audio quality feedback, minimal jargon. Non-technical users must produce clinically usable samples.
The voice donor flow is browser-based and works on any device. Real-time feedback on audio quality, an explicit consent step tied to the patient's name, and a clear path to remove the recording at any time.
Transparent AI onboarding: uncanny valley mitigation
Explicitly stated the voice is AI-generated. Preview-before-use. Full user control over activation. Trust > novelty. In healthcare, perceived deception destroys adoption.
Emotional framing
Positioned as connection-enhancing, not gimmicky. Reinforced dignity and familiarity throughout.
Key Screens
System Diagram
Collaboration
Worked across AI engineering (voice synthesis constraints), clinical leadership (therapeutic validity), legal & compliance (consent and data usage), and product marketing (positioning differentiation).
Navigated data privacy concerns, emotional sensitivity, technical constraints, and cross-platform deployment.
Outcomes & Impact
- Became a primary differentiator for Lace Pro
- Strengthened positioning as AI-powered auditory rehabilitation
- Increased provider excitement and demo conversion
- Elevated perceived innovation maturity of the platform
- Reinforced premium positioning of the product suite
Leadership
- High-complexity AI initiative leadership
- Ethical AI UX judgment in regulated healthcare
- Multi-stakeholder orchestration (AI, clinical, legal, product, marketing)
- Scalable personalization systems
- Emotional UX in clinical contexts