Sanvia
Sanvia
Sanvia
Sanvia
AI Health Platform
AI Health Platform
Personalized Answers
Personalized Answers
Secure Information
Secure Information
Details
Category
AI
Health
Type
Product Design
Team
BU Spark!
Duration
5 months
Sanvia is a health knowledge platform that uses AI to offer clear, personalized answers to your health questions without privacy concerns.
Details
Type
Product Design
Case Study
Role
Product Designer
Researcher
Timeline
Spring 2024
BU Spark Innovation Program
Category
AI
Health
Role
Type
Type
Product Designer
Product Design
Product Design
Team
BU Spark!
Timeline
Duration
Duration
Spring 2025
5 Months
5 Months
Sanvia is a health knowledge platform that uses AI to offer clear, personalized answers to your health questions without privacy concerns.
Sanvia is a health knowledge platform that uses AI to offer clear, personalized answers to your health questions without privacy concerns.
Sanvia is a health knowledge platform that uses AI to offer clear, personalized answers to your health questions without privacy concerns.
[01]
problem
Looking up health questions online is stressful and often misleading.
I need a simple, safe way to get health answers online!
Articles from WebMD or Healthline are too general, full of jargon, and often alarming (“You might have cancer”).
People spend hours refining searches but still struggle to find relevant insights for their specific conditions.
Even conversational AI tools like ChatGPT can feel vague and raise privacy concerns with sensitive data.
[01]
problem
Looking up health questions online is stressful and often misleading.
Looking up health questions online is stressful and often misleading.
I need a simple, safe way to get health answers online
I need a simple, safe way to get health answers online
I need a simple, safe way to get health answers online
I need a simple, safe way to get health answers online
Articles from WebMD or Healthline are too general, full of jargon, and often alarming (“You might have cancer”).
Articles from WebMD or Healthline are too general, full of jargon, and often alarming (“You might have cancer”).
Articles from WebMD or Healthline are too general, full of jargon, and often alarming (“You might have cancer”).
Articles from WebMD or Healthline are too general, full of jargon, and often alarming (“You might have cancer”).
People spend hours refining searches but still struggle to find relevant insights for their specific conditions.
People spend hours refining searches but still struggle to find relevant insights for their specific conditions.
People spend hours refining searches but still struggle to find relevant insights for their specific conditions.
People spend hours refining searches but still struggle to find relevant insights for their specific conditions.
Even conversational AI tools like ChatGPT can feel vague and raise privacy concerns with sensitive data.
Even conversational AI tools like ChatGPT can feel vague and raise privacy concerns with sensitive data.
Even conversational AI tools like ChatGPT can feel vague and raise privacy concerns with sensitive data.
Even conversational AI tools like ChatGPT can feel vague and raise privacy concerns with sensitive data.
[02]
discovery phase
Creating a Customer Profile
Early Evangelists Checklist
We identified two main audiences: health-conscious patients and chronic disease patients. Health-conscious patients look for guidance on lifestyle, habits, and preventative care. Chronic patients need more support understanding treatments, managing logistics, and keeping track of their health information.
Understanding the User
User
Job (I'm trying to…)
Pains
Gains
Health-concerned individuals (not always diagnosed patients) seeking clarity on symptoms or conditions
Understand the cause and severity of symptoms
Overwhelmed by vast, messy, and conflicting information online
Clear, relevant, and personalized health information
Patients feeling overwhelmed or unsatisfied with current care
Decide on next steps: doctor visit, OTC medication, self-monitoring, emergency care
Hard to find answers that feel relevant to personal context and manual data entry feels tedious
Reduced time and effort in finding and filtering medical advice
Chronic condition patients looking for better self-care guidance
Access relevant resources without excessive searching/filtering
Limited resources: time, money, energy, insurance coverage, etc.
Stress relief through actionable insights and reassurance
Health-conscious individuals seeking preventative measures
See if others share similar conditions for reassurance
Lack of next-step clarity or inspiration
Ability to keep health data private and under personal control
People searching for new solutions after medical advice fell short
Quickly check if symptoms are related to a known condition
Reliance on unreliable anecdotal advice (ex. Reddit)
Access to trusted resources without overwhelming effort
Persona:
Digital Dork Daniel (20)
Bio
STEM undergrad, power user of AI assistants for school and daily life. Active in online forums (Reddit, Discord, Hacker News), curious about biohacking and self-optimization.
Needs
Fast, personalized, AI-driven health insights
Clarity without jargon
Trustworthy, up-to-date information
Pains
Distrusts generic or outdated health advice
Overwhelmed by technical language
Concerned about privacy/security
Preferences
Clean, intuitive UI/UX
Quick, automated support
Data-backed, credible tools
Persona:
Digital Dork Daniel (20)
Bio
STEM undergrad, power user of AI assistants for school and daily life. Active in online forums (Reddit, Discord, Hacker News), curious about biohacking and self-optimization.
Needs
Fast, personalized, AI-driven health insights
Clarity without jargon
Trustworthy, up-to-date information
Pains
Distrusts generic or outdated health advice
Overwhelmed by technical language
Concerned about privacy/security
Preferences
Clean, intuitive UI/UX
Quick, automated support
Data-backed, credible tools
Persona:
Bio
STEM undergrad, power user of AI assistants for school and daily life. Active in online forums (Reddit, Discord, Hacker News), curious about biohacking and self-optimization.
Needs
Fast, personalized, AI-driven health insights
Clarity without jargon
Trustworthy, up-to-date information
Pains
Distrusts generic or outdated health advice
Overwhelmed by technical language
Concerned about privacy/security
Preferences
Clean, intuitive UI/UX
Quick, automated support
Data-backed, credible tools
[02]
discovery phase
Creating a
Customer Profile
Early Evangelists Checklist
Has a chronic disease
Frequent user of digital media & internet
Paranoid about their health at least once a month
Use NLP like ChatGPT / Open to use AI for health
Digital Dork Daniel (20)
Bio
STEM undergrad, power user of AI assistants for school and daily life. Active in online forums (Reddit, Discord, Hacker News), curious about biohacking and self-optimization.
Needs
Fast, personalized, AI-driven health insights
Clarity without jargon
Trustworthy, up-to-date information
Pains
Distrusts generic or outdated health advice
Overwhelmed by technical language
Concerned about privacy/security
Preferences
Clean, intuitive UI/UX
Quick, automated support
Data-backed, credible tools
We identified two main audiences: health-conscious patients and chronic disease patients. Health-conscious patients look for guidance on lifestyle, habits, and preventative care. Chronic patients need more support understanding treatments, managing logistics, and keeping track of their health information.
[03]
research
Key Findings and Insights
Approach
Approach
1
1
13 user interviews with health-conscious individuals and chronic condition patients
13 user interviews with health-conscious individuals and chronic condition patients
2
2
Surveys and street interviews
(18–60 years old)
Surveys and street interviews (18–60 years old)
3
3
Competitive analysis of major platforms: WebMD, Healthline, ChatGPT, Ada Health, Sensely, Buoy, Mayo Clinic, Hippocratic AI
Competitive analysis of major platforms: WebMD, Healthline, ChatGPT, Ada Health, Sensely, Buoy, Mayo Clinic, Hippocratic AI
Findings
88% searched online before doctor visits
85% found information confusing, unverifiable, or too general
Privacy concerns were common, but users were open to AI if trust was established
Pain points: overwhelming info, repetitive advice, and lack of personalization
Desire gains: clarity, reassurance, speed, and context-aware answers
Competitive Analysis
Competitive Analysis
Credible but Impersonal
WebMD, Healthline, and Mayo Clinic are trusted and medically verified sources, but their interfaces can feel overwhelming and crowded with ads. The content is often filled with jargon, making it hard to digest, and the information tends to be too generic to address personal health needs.
Conversational but Unverified
AI chatbots like ChatGPT are easy to use, conversational, and adaptable in tone, but they are not health-specific and can be prone to inaccuracies or vague answers. They also lack medical oversight, and data privacy concerns make many users hesitant to share sensitive health information.
Health-Specific AI Tools
Hippocratic AI is accurate, HIPAA-compliant, and trusted in clinical settings, but it is designed primarily for healthcare professionals rather than everyday consumers. Platforms like Buoy Health, Ada Health, and Sensely offer personalization and symptom checking, yet they often require lengthy questionnaires, have dated interfaces, or focus on narrow scenarios, which can create friction and reduce usability.
Credible but Impersonal
WebMD, Healthline, and Mayo Clinic are trusted and medically verified sources, but their interfaces can feel overwhelming and crowded with ads. The content is often filled with jargon, making it hard to digest, and the information tends to be too generic to address personal health needs.
Conversational but Unverified
AI chatbots like ChatGPT are easy to use, conversational, and adaptable in tone, but they are not health-specific and can be prone to inaccuracies or vague answers. They also lack medical oversight, and data privacy concerns make many users hesitant to share sensitive health information.
Health-Specific AI Tools
Hippocratic AI is accurate, HIPAA-compliant, and trusted in clinical settings, but it is designed primarily for healthcare professionals rather than everyday consumers. Platforms like Buoy Health, Ada Health, and Sensely offer personalization and symptom checking, yet they often require lengthy questionnaires, have dated interfaces, or focus on narrow scenarios, which can create friction and reduce usability.
Credible but Impersonal
WebMD, Healthline, and Mayo Clinic are trusted and medically verified sources, but their interfaces can feel overwhelming and crowded with ads. The content is often filled with jargon, making it hard to digest, and the information tends to be too generic to address personal health needs.
Conversational but Unverified
AI chatbots like ChatGPT are easy to use, conversational, and adaptable in tone, but they are not health-specific and can be prone to inaccuracies or vague answers. They also lack medical oversight, and data privacy concerns make many users hesitant to share sensitive health information.
Health-Specific AI Tools
Hippocratic AI is accurate, HIPAA-compliant, and trusted in clinical settings, but it is designed primarily for healthcare professionals rather than everyday consumers. Platforms like Buoy Health, Ada Health, and Sensely offer personalization and symptom checking, yet they often require lengthy questionnaires, have dated interfaces, or focus on narrow scenarios, which can create friction and reduce usability.
Credible but Impersonal
WebMD, Healthline, and Mayo Clinic are trusted and medically verified sources, but their interfaces can feel overwhelming and crowded with ads. The content is often filled with jargon, making it hard to digest, and the information tends to be too generic to address personal health needs.
Conversational but Unverified
AI chatbots like ChatGPT are easy to use, conversational, and adaptable in tone, but they are not health-specific and can be prone to inaccuracies or vague answers. They also lack medical oversight, and data privacy concerns make many users hesitant to share sensitive health information.
Health-Specific AI Tools
Hippocratic AI is accurate, HIPAA-compliant, and trusted in clinical settings, but it is designed primarily for healthcare professionals rather than everyday consumers. Platforms like Buoy Health, Ada Health, and Sensely offer personalization and symptom checking, yet they often require lengthy questionnaires, have dated interfaces, or focus on narrow scenarios, which can create friction and reduce usability.
How might we utilize the gaps in existing solutions to make Sanvia stand out?
How might we utilize the gaps in existing solutions to make Sanvia stand out?
How might we utilize the gaps in existing solutions to make Sanvia stand out?
Gap
No existing platform delivers personalized, medically credible, conversational guidance in an ad-free, intuitive experience that also respects privacy and gives clear next steps for users
Gap
No existing platform delivers personalized, medically credible, conversational guidance in an ad-free, intuitive experience that also respects privacy and gives clear next steps for users
Opportunity Space
Sanvia bridges this gap by combining:
Credibility: sourcing from verified medical resources (WebMD, NIH)
Personalization: integrating personal health data, wearable insights, and symptom history
Conversational clarity: guiding users in natural language with follow-up prompts
Privacy & trust: transparent data handling and HIPAA-conscious design
Actionable next steps: helping users navigate decisions with confidence, not fear
Opportunity Space
Sanvia bridges this gap by combining:
Credibility: sourcing from verified medical resources (WebMD, NIH)
Personalization: integrating personal health data, wearable insights, and symptom history
Conversational clarity: guiding users in natural language with follow-up prompts
Privacy & trust: transparent data handling and HIPAA-conscious design
Actionable next steps: helping users navigate decisions with confidence, not fear
Findings
88% searched online before doctor visits
85% found information confusing, unverifiable, or too general
Privacy concerns were common, but users were open to AI if trust was established
Pain points: overwhelming info, repetitive advice, and lack of personalization
Desire gains: clarity, reassurance, speed, and context-aware answers
[04]
design process
Design and Pivot
Early
Moodboard
Early Concepts
Our early builds emphasized personalization and resourcefulness. This was reflected in a light mode that emphasized key words like educational and trustworthy.
The design was meant to be minimal and intuitive; suggesting that health should be easy to navigate.
Usability Testing
Our team continued conducting interviews showcasing our low-fidelity wireframes and initial developmental team prototypes.
We completed 15 interviews, aimed at early evangelists and previous interview participants.
In our remaining sprints, we conducted a usability interview after we designed and implemented a new feature each week, completing 7 more interviews.
Branding
Logo
Inspired by the Latin roots of "san" for health and "via" for path, Sanvia was named to represent where health meets your personal pathway.
This was represented in our logo, showing a path shaped into the letterform S.
Privacy by Design
After user interviews, we doubled down on de-identification and user control and realized our differentiator was privacy.
We reworked our platform around end-to-end encryption and full data ownership, implementing a dark mode interface to visually signal security.
Branding
Logo
Inspired by the Latin roots of "san" for health and "via" for path, Sanvia was named to represent where health meets your personal pathway.
This was represented in our logo, showing a path shaped into the letterform S.
Early Iterations
In today’s world, where people are often wary of AI, it’s crucial that design choices foster trust.
My approach focuses on creating an empathetic connection between Sanvia and the user, ensuring the experience feels approachable, safe, and human-centered.
As part of this focus on security, we shifted from a light mode to a dark mode interface. The darker palette was intentional: it visually reinforces privacy and safety while keeping the design minimal and intuitive, suggesting that health should be easy to navigate.
The typeface and colors balances clarity and emotion: a clean typeface ensures approachability, while a sleek, futuristic accent font signals modernity. Shades of blue convey trust and communication, paired with purple to evoke healing, wisdom, and consciousness.
This together creates a visual language that feels both modern and emotionally supportive in the midst of the overwhelming healthcare landscape.
Early Iterations
[04]
design process
Design Decisions
Early
Moodboard
In today's world, where people are often wary of AI, it’s crucial that design choices foster trust.
My approach focuses on creating an empathetic connection between Sanvia and the user, ensuring the experience feels approachable, safe, and human-centered.
In today's world, where people are often wary of AI, it’s crucial that design choices foster trust.
My approach focuses on creating an empathetic connection between Sanvia and the user, ensuring the experience feels approachable, safe, and human-centered.
Early Concepts
Our early builds emphasized personalization and resourcefulness. This was reflected in a light mode that emphasized key words like educational and trustworthy.
Usability Testing
Our team continued conducting interviews showcasing our low-fidelity wireframes and initial developmental team prototypes.
We completed 15 interviews, aimed at early evangelists and previous interview participants.
In our remaining sprints, we conducted a usability interview after we designed and implemented a new feature each week, completing 7 more interviews.
Privacy by Design
After user interviews, we doubled down on de-identification and user control and realized our differentiator was privacy.
We built the platform around end-to-end encryption and full data ownership, ensuring users feel safe sharing sensitive health information.
As part of this focus on security, we shifted from a light mode to a dark mode interface. The darker palette was intentional: it visually reinforces privacy and safety while keeping the design minimal and intuitive, suggesting that health should be easy to navigate.
As part of this focus on security, we shifted from a light mode to a dark mode interface. The darker palette was intentional: it visually reinforces privacy and safety while keeping the design minimal and intuitive, suggesting that health should be easy to navigate.
Branding
Logo
Inspired by the Latin roots of "san" for health and "via" for path, Sanvia was named to represent where health meets your personal pathway.
This was represented in our logo, showing a path shaped into the letterform S.
Early Iterations
The typeface and colors balances clarity and emotion: a clean typeface ensures approachability, while a sleek, futuristic accent font signals modernity. Shades of blue convey trust and communication, paired with purple to evoke healing, wisdom, and consciousness.
This together creates a visual language that feels both modern and emotionally supportive in the midst of the overwhelming healthcare landscape.
The typeface and colors balances clarity and emotion: a clean typeface ensures approachability, while a sleek, futuristic accent font signals modernity. Shades of blue convey trust and communication, paired with purple to evoke healing, wisdom, and consciousness.
This together creates a visual language that feels both modern and emotionally supportive in the midst of the overwhelming healthcare landscape.
[05]
problem-solving
Solution Features
Personalized search that uses your documents and health data
Context-aware AI tuned specifically for health questions
Explainable answers with the ability to ask clarifying follow-ups
Strict privacy with end-to-end encryption and user-controlled data
Onboarding Survey
Initial health profile
“Summarize my last health checkup report.”
Document Upload
Lab results, doctor’s notes, diagnostic reports
Ask questions like “Summarize my last health checkup report” or “Is my sleep improving?” to receive responses based on uploaded data
Personalized Profile & Wearable Integration
Optimized contextual response and connect health tracking platforms to pull in relevant metrics like sleep and activity
Medical Chat History
Revisit past chats to integrate and upload updated medical information
Medical Bibliography
Check the original sources of the information provided
[05]
problem-solving
Solution Features
Personalized search that uses your documents and health data
Context-aware AI tuned specifically for health questions
Explainable answers with the ability to ask clarifying follow-ups
Strict privacy with end-to-end encryption and user-controlled data
Onboarding Survey
Initial health profile
“Summarize my last
health checkup report.”
Document Upload
Lab results, doctor’s notes, diagnostic reports
Ask questions like “Summarize my last health checkup report” or “Is my sleep improving?” to receive responses based on uploaded data
Personalized Profile & Wearable Integration
Optimized contextual response and connect health tracking platforms to pull in relevant metrics like sleep and activity
Medical Chat History
Revisit past chats to integrate and upload updated medical information
Medical Chat History
Revisit past chats to integrate and upload updated medical information
Medical Bibliography
Check the original sources of the information provided
Landing
Landing
Chat Interface
Chat Interface
[05]
problem-solving
Solution Features
Personalized search that uses your documents and health data
Context-aware AI tuned specifically for health questions
Explainable answers with the ability to ask clarifying follow-ups
Strict privacy with end-to-end encryption and user-controlled data
Onboarding Survey
Initial health profile
Onboarding surveys can feel tedious or like a waste of time, especially when they aren’t engaging. To maximize retention, we added interactive animations with moving gradients and streamlined the questions to only the essential information, such as age and sex.
“Summarize my last health checkup report.”
Document Upload
Lab results, doctor’s notes, diagnostic reports
Ask questions like “Summarize my last health checkup report” or “Is my sleep improving?” to receive responses based on uploaded data
Personalized Profile & Wearable Integration
Optimized contextual response and connect health tracking platforms to pull in relevant metrics like sleep and activity
Medical Chat History
Revisit past chats to integrate and upload updated medical information
Medical Bibliography
Check the original sources of the information provided
[05]
problem-solving
Solution Features
Personalized search that uses your documents and health data
Context-aware AI tuned specifically for health questions
Explainable answers with the ability to ask clarifying follow-ups
Strict privacy with end-to-end encryption and user-controlled data
Onboarding Survey
Initial health profile
“Summarize my last
health checkup report.”
Document Upload
Lab results, doctor’s notes, diagnostic reports
Ask questions like “Summarize my last health checkup report” or “Is my sleep improving?” to receive responses based on uploaded data
Personalized Profile & Wearable Integration
Optimized contextual response and connect health tracking platforms to pull in relevant metrics like sleep and activity
Medical Chat History
Revisit past chats to integrate and upload updated medical information
Landing
Landing
Chat Interface
Chat Interface
Landing
Chat Interface
[06]
presentation
Demo Day
Presentation Day
At the end of the Spark! Innovation Program, our team presented Sanvia on Demo Day to the Boston University Faculty for Computing & Data Sciences with a presentation and on-display Eposters.
[06]
presentation
Demo Day
Presentation Day
At the end of the Spark! Innovation Program, our team presented Sanvia on Demo Day to the Boston University Faculty for Computing & Data Sciences with a presentation and on-display Eposters.
Eposter
[06]
reflection
Looking Ahead
Ethics Assessment
Ethics Assessment
Sanvia held consultations with the Ethics and Law department at Boston University to address any concerns in those areas.
Key risk areas included privacy and security, which we addressed through end-to-end encryption, de-identification, and full user control over data. We mitigated regulatory risk by clearly positioning Sanvia as an educational medical knowledge platform, not a diagnostic tool, to remain within the FDA’s low-risk zone.
Sanvia held consultations with the Ethics and Law department at Boston University to address any concerns in those areas.
Key risk areas included privacy and security, which we addressed through end-to-end encryption, de-identification, and full user control over data. We mitigated regulatory risk by clearly positioning Sanvia as an educational medical knowledge platform, not a diagnostic tool, to remain within the FDA’s low-risk zone.
Future Plans
Future Plans
Our future plans include fully implementing the design prototype and refining the UI.
The development team future plans include: expanding data integrations (Apple Health, EPIC), launching mobile apps and web support, and offering advanced features like licensed knowledge mining, goal-based health insights, 3D visualizations, and anonymized comparison programs.
Our future plans include fully implementing the design prototype and refining the UI.
The development team future plans include: expanding data integrations (Apple Health, EPIC), launching mobile apps and web support, and offering advanced features like licensed knowledge mining, goal-based health insights, 3D visualizations, and anonymized comparison programs.
Takeaways
Takeaways
Personalization must feel transparent to avoid overwhelming the audience.
Visual cues matter: dark mode and data controls became signals of safety.
Designing for health and AI requires constant balancing of education and caution.
Personalization must feel transparent to avoid overwhelming the audience.
Visual cues matter: dark mode and data controls became signals of safety.
Designing for health and AI requires constant balancing of education and caution.
Landing