Olivia Health is an AI-powered app launched in 2024 from Tempus that aims to support patients and caregivers in navigating care. Another designer and myself conducted a short, high-impact collaboration to audit the product and identify new opportunities to make Olivia more proactive and useful.
The goal was to expand Olivia beyond passive data storage into an intelligent assistant that guides users through their care journey – especially in high-stakes medical moments.
After reviewing user feedback of the Olivia app, we conducted a UX audit of the platform. These revealed pain points around unclear value, lack of trust in data handling, and minimal proactive guidance from the AI.
We also identified broader healthcare insights – such as the caregiver burden during appointments – that shaped our approach.
Our audit surfaced four key themes, each revealing distinct opportunity areas that ultimately informed our guiding How Might We question.
How might Olivia guide patients, caregivers, and health providers through the entire care journey so they can better share information, make decisions, and understand the value of treatments?
How might Olivia guide patients, caregivers, and health providers through the entire care journey so they can better share information, make decisions, and understand the value of treatments?
I prototyped three distinct user flows focused on critical moments in the doctor-patient-caregiver relationship: before, during, and after the appointment. Each prototype was designed to make Olivia feel more intelligent, helpful, and human while offering timely support and making complex medical interactions more understandable.
My design decisions were guided by principles of clarity, timing, and trust, ensuring that Olivia would show up at the right moment, in the right tone, and with the right level of support.
Pre-Appointment
I designed features that allow Olivia to proactively generate personalized questions based on the user’s medical history, symptom patterns, and upcoming appointment context.
In-Appointment
I designed a real-time visit mode with a built-in session timer, smart suggestions, and a conversational UI where Olivia could “chime in” with clarifying or follow-up questions.
Post-Appointment
I prototyped a concise, AI-generated summary of the visit that included diagnoses, next steps, and personalized follow-up reminders.
Tempus has since implemented many of the features we proposed, including the appointment prep checklist, AI-generated question prompts, in-visit note-taking tools, and visit summaries with chat-based clarification. These features elevated Olivia from a passive logbook to a proactive companion.
While we didn’t ship directly, our concepts informed the next product release and helped Tempus reframe Olivia’s purpose around trust, personalization, and real-time support.
I also led the framing of the overall product narrative and themes, contributed to the strategy for simplifying onboarding, and shaped storytelling for the design audit presentation.
Designing for trust in AI requires clarity, empathy, and visible value – especially in healthcare. If I did this again, I’d bring in patients or caregivers earlier to co-create the prep and reflection tools. Still, it was inspiring to see our work implemented so quickly, proving how fast and suggestive work can still drive real-world impact.
