Overview

Venera AI is a health technology company that uses advanced AI to turn complex medical data into clear, evidence-based insights. Originally designed for healthcare professionals, our platform leverages natural language processing and machine learning to analyze patient medical reports and clinical databases, helping clinicians make faster, data-driven decisions.

When I joined in early 2024, the company was expanding beyond hospitals and clinics to empower chronic patients directly, shifting from a B2B to a B2C model. The goal was to identify daily tasks that chronic patients, who manage their health independently between doctor visits, struggle with most, and design a product that could empathize with their daily challenges and offer meaningful support. Before my arrival, the team had already begun exploring this direction, evaluating how the existing clinical AI-powered technology could be adapted to meet the needs of our new users.

My Role: I led the end-to-end design process from initial research to launch-ready product. I conducted stakeholder and user interviews, facilitated cross-functional workshops, and identified the advanced AI-powered features that would drive subscription revenue through our freemium model. I built a scalable Figma design system with reusable components and design tokens, increasing design velocity by 20%. I also validated and refined key design decisions through multiple rounds of iterative prototyping and usability testing, resulting in a 26% increase in user satisfaction.

What did I discover early?

Through initial research, I learned three key insights. First, most chronic patients don’t manage their health alone, they often rely on caregivers and healthcare professionals to help with their daily routines and treatment decisions. These findings made it clear that if we wanted to build a product that patients would actually trust and use, it needed to reflect the perspectives of not only patients but also caregivers and healthcare professionals. Second, the reason most patients stop using health apps or other digital platforms, or hesitate to take AI-powered advice is that these tools lack clinical validation. Without clinical validation, both patients and healthcare professionals question their accuracy and credibility. Finally, I found that patients primarily use their phones to manage daily tasks, such as tracking metrics or checking patient portals. Therefore, we decided that our new product should be a mobile app, rather than a web platform like the product for clinicians.

From that point, I conducted interviews with twenty patients and caregivers about their daily challenges and feature expectations, followed by more interviews with ten healthcare professionals who would provide their clinical perspectives on the insights from previous patient interviews. This approach ensured all voices would be reflected in our final design.

What did I learn from research?

Both patients and healthcare professionals agreed that diet is one of the most difficult parts of managing chronic condition. Patients often struggle to make healthy food choices, especially when dining out or celebrating with friends and colleagues. Many stop following their diets once they start feeling better, unaware of the long-term risks. Healthcare professionals highlighted that progress comes from small, consistent changes rather than strict restrictions or sudden changes. Building confidence through achievable goals and showing how diet connects to their overall well-being can help patients stay on track.

Patients and caregivers wanted simple, clear information about what causes their symptoms and how their condition might progress. Many patients felt lost after brief or overly complex medical consultations. Healthcare professionals, on the other hand, focused on managing expectations, emphasizing that chronic conditions cannot be “cured”, they must be managed over time. They noted that both patients and their caregivers must learn to accept and manage the condition, rather than seek magic cures from unverified sources. This difference revealed an opportunity to use our existing AI technology and clinical databases to help patients better understand their condition, set realistic expectations, and feel more in control of their health journey.

Patients and caregivers were often overwhelmed by the number of medication brands and conflicting instructions from multiple doctors. Many were hesitant to use their prescribed medications out of fear that they might experienced dangerous side effects from taking a wrong combination (especially from patients with multiple chronic conditions). Healthcare professionals worried that this avoidance, which was often fueled by misinformation, could prevent effective treatment. Both groups saw the need for better communication and personalized guidance to help patients follow their treatment plans confidently.

Patients described feelings of grief and helplessness after their diagnosis, struggling to adjust to new limitations and the loss of their daily routines. Caregivers also carried a heavy emotional burden, often feeling anxious or guilty about their role in treatment decisions. Many questioned whether they were doing enough or making the right choices for their loved ones, while also facing pressure from family, friends, and even other clinicians. These experiences highlight the need to design features that provide reassurance and emotional support. Our product would fail if it treated users as data points rather than recognizing the human emotions behind their experiences.

How did I decide which product ideas to pursue?

With a wide range of ideas from stakeholder interviews, user requests, and research insights, I shifted from exploring possibilities to deciding which ideas to pursue. To do this, I held a prioritization workshop with a cross-functional team that included business advocates (our company founders), a user advocate (myself), developers and ML engineers, and a product manager. Each participant received dots to vote on ideas, value dots for business and user advocates, and complexity dots for developers, based on how impactful or effortful we believed each idea would be. This collaborative exercise helped us quickly uncover which ideas offered the most value for the least complexity.

We then mapped all ideas onto a Value–Complexity matrix, where those in the high-value and low-complexity quadrant, our Quick Wins, became the starting point for MVP roadmap. These features included:

• Chatbot

• Meal logging and recommendations

• Social event planning

• Medication tracking and conflict detection

• Medical report summarization

How did I validate my design decisions?

After the internal concept review and several rounds of low-fidelity iteration, I created a high-fidelity interactive prototype to test three critical design decisions for the MVP release with 12 participants.

Testing Method: Moderated usability testing with think-aloud protocol. Success was measured by task completion without requiring Level 2 or Level 3 prompts (Level 2 = hints/clarification; Level 3 = direct instruction or user gives up).

The testing focused on three key objectives:

1. Validating home screen’s content: Determining whether users prefer seeing and getting quick access to their medication tracking, meal recommendation, or medical report summarization.

2. Testing manual logging: Assessing how willing users are to manually log meals, medications, and medical reports, and whether the design feels simple and intuitive enough to do so without confusion.

3. Evaluating feature usability: Examining how users understand and trust the AI-powered features and identifying any concerns that might affect their willingness to authorize data access.