Elsa AI
An AI-first assistant making egg-freezing process clear, empathetic + accessible
Designed as a conceptual product with an emphasis on AI UX principles to improve my AI UX design skills
How can we using AI advancements to create an egg freezing assistant to improve women's fertility success rates?
Design Process
Understand pain points > establish requirements > ideate AI solutions > lofi > Lovaable > refine lovable
Role
As Senior UX Designer, this was a passion project to get more comfortable using AI within a UX context
Goal
Design an AI-first assistant that restores confidence, clarifies decision points, and humanizes the egg-freezing experience.
Outcome
Built a functional AI prototype in one month using Figma, Lovable, and ChatGPT, enhancing human-AI collaboration and user empowerment
Project Details
Meet Mia Castillo
34, tech-savvy, emotionally cautious - juggles anxiety and urgency around fertility choices
“I want to make sure I make the right decisions before it’s too late”
Navigating Egg Freezing
“I want to feel like I’m not falling behind in life just because I haven’t figured everything out yet. I need support that gets me—something that helps me make smart decisions without scaring me.”
Tech Behaviors
Tracks cycle using Clue app
Uses Headspace for guided meditation
Shops online, values personalized recommendations
Comfortable with chatbots if they feel intelligent and friendly
Fragmented, clinical info
Overwhelming number of choices
No real-time guidance
Where Mia is Struggling the Most
This Wellness Tracker is an example of all 5 principles in action, seemlessly building trust and transparency for the user
AI UX Principles for Mia's Needs
Clear timeline of steps
Understand trade-offs (freeze now vs. later)
Emotional reassurance
Which clinic and timeline
How many cycles to pursue
Balancing costs and health risks
Anxious, isolated, uncertain
Craving agency and calm
Relief when supported
I designed Elsa AI to solve the challenges I faced as an egg freezing patient and to advance my leadership in AI product design
Legal + Regulatory
Define which regions it would work in
Data Collection
Only collect the minimum data needed. Use de-identified or pseudonymized data where possible. Be explicit with users about what is collected, how used.
Consent + Transparency
Create a proper informed consent flow before collecting sensitive info. Provide privacy notice, terms of service. Users can control / revoke consent at any time
Data Storage + Transit
Use encryption at rest and in transit
User Authorization
Secure user login, possibly 2-factor
UX / UI Elements
Privacy settings in UI; showing disclaimers; making transparent what AI is doing; letting users see how recommendations were generated; letting users choose how much personal data to share.
Data Security, Privacy + Compliance
AI Fluency in Practice
A big part of working with the AI tools was leading and directing the AI outputs with prompts. Below are some examples of: edited/rejected outputs + human-centered refinements
Key AI UX Features
Decision Stimulation Tool
Research
Fertility decisions are too abstract, hard to compare - users struggle with trade-offs, like: cost, age and success rates
Design Strategy
Enables calculators, such as “freeze now vs. later,” to provide large sets of data with easy to comprehend visuals
Smart Personalized Dashboard
Research
Users feel overwhelmed by too many choices + want clear guidance
Design Strategy
Create a first-impression that streamlines next steps, reduces cognitive load, leading to fostered confidence
AI Nurse Chat with Human Escalation
Research
Users wished they had someone to consult with when they felt uncertain, esp. after hours when offices are closed
Design Strategy
Implement an AI nurse chat bot that can be accessed 24/7 to relieve anxiety and offer support - backed when needed by real support
Emotionally Adaptive UI
Research
When data feels "too clinical" users disengage
Decision Strategy
Balance clinical info with empathetic presentation to build engagement + trust
How I'd Measure Success if Live
Final Design
Lessons Learned
Designing Systems
Designing AI is really just building systems that are: predictable, emotional + easily explainable
Prioritize Patient's Readiness
Patient's emotional readiness is as important as clinical data in health decisions - patients need to feel grounded + secure
Trust is Built
Trust in an AI filled world is built through: transparency, optionality + progressively ("slow" is the fastest way)