Anatomy of Personalized Telehealth

I built the foundational AI-based telehealth tool for Prana. This has been aiding in their fundraising efforts and helped in onboarding double the users from previous quarter.

Team: Me (Design, Research), CEO, 3 Engineers

Duration: 1 month Sprint

01. CONTEXT

Prana is a personal health platform that helps users understand their lab results, biomarker data, overall goal-based wellness through GenAI. It bridges the gap between raw clinical data and real-world action by combining AI-driven summaries with optional clinician consults, all within a HIPAA-compliant, mobile-first experience.

Goals

Translate user-generated health data into clinician-ready context: Bridge Prana’s rich personal data (labs, symptoms, goals) into structured, actionable inputs for clinical decision-making.

Drive meaningful action toward longer, healthier lives: Make it easy for users to move from insight to care, improving early intervention, adherence, and health outcomes.

02. RESEARCH

10 User Interviews

with individuals who recently received low B12 results via Prana. We explored how they interpreted the insight, what action they took (if any), and what kind of clinical intervention felt accessible or overwhelming.

4 clinician interviews

(Zuckerberg General) focused on how providers manage Prana patients and what kind of context or tools would improve consult efficiency.

3 PSYCHOLOGIST INTERVIEWS

were triggered by unexpected patterns around fatigue, uncertainty, and health anxiety in user responses, highlighting the emotional layer of decision paralysis.

Research Results

60% of users turned to Reddit or other forums after receiving their B12 results and one user quotes their "Fear of overthinking"

Clinicians shared that managing Prana users through manual Excel tracking was becoming unsustainable

Psychologists added a key insight: the act of scheduling itself can be a moment of empowerment, helping users shift from passive worry to proactive care

03. DESIGN SPECS

User Flow

After mapping out the flow, we defined the core values, and prioritized release plan.

+ Design for Trust

+ Be Forgiving by Default

+ Show Immediate Value

+ Consider PHI compliance

03. DESIGN

I built a comprehesive app flow to understand the distinct scenarios and how the AI assistant would manifest in each of them. Below are the scenariors with the most time reduction due to AI

Notification on the Homepage

Intutive and motivating scheduler

Quick chat and coversation with provider for personalized care

Iterative Prototyping

I rapidly iterated on the scheduling and prep flow through three rounds of user interviews and prototype reviews, testing with both patients and clinicians. Each round helped refine how we framed discussion points, surfaced lab data, and handled consent. One unexpected insight was the desire for lighter, asynchronous follow-up—users didn’t always want to book a full consult. In response, I designed and tested a secure chat feature, allowing patients to ask quick questions post-visit and continue care without friction. This addition helped shift the experience from episodic to relational.

Quick chat and coversation with provider for personalized care

Also check out..