top of page
Fern
A next generation chronic pain management app (Mobile & Web) that combines the power of conversational AI with the expertise of a human health coach. With a holistic approach to care, users can expect a personalized and comprehensive care plan that empowers them to take control of their pain and improve their overall quality of life.
Main opportunity: Reinvent a digital therapeutic for people with chronic pain, by leveraging generative AI and human coach assistants in order to deliver warm, personalized care.
Outcomes that deliver
2.5 x
Greater likelihood of a clinical reduction in disability.
20 %
Improvement in pain catastrophizing.
97 %
Of Fern users would refer a family member or friend.
Illustration by /in/morgan-lines
Illustration by /in/morgan-lines
Derived from SMART goal setting techniques, Fern helps users set and track goals. We took the clinical vision of having "nested" goals and quick wins and broke it down into an activity that makes sense. Through several iterations and user testing we were able to design a flow that was able to take a complex idea and make it simple and intuitive.
Goal Setting flow:
-
Overview video to set expectations for the activity
-
Select a focus area
-
Work with Fern's AI to understand what's realistic and achievable for you based on your focus area. Establish a starting point.
-
Set a Goal and Fern will help make sure it is SMART
-
Break it down into quick wins you can do this week
-
Set Reminders
Addressing the Issue
Fern is a live app that delivers chronic pain management content to users, however the majority of people do not make it through intake. Only a select few that make it through intake will actually find enjoyment in using the program. However, the people that do make it through the 8 week program know how much potential the app has.
Main areas for improvement:
-
Improve intake flow - too many questions, takes too long to complete, leaves user wondering what the app is and how it will help them.
-
Desirability & Engagement - most content is reading, no rewards, no motivation to come back besides self driven progress.
-
Expectation Setting
-
Assessments
-
Preferred Voice and Tone
-
Understanding each users "why" behind wanting to feel better
-
-
Personalization - most users get the same content no matter what they selected as their pain area. The program itself is not tailored to each user. No feedback loop in giving my pain info and seeing it make a difference in the program.
Goal: Create a digital therapeutic application that understands the day in the life of chronic pain, delivers care with warmth and 24/7 assistance from generative AI and human coach assistance.
Voice & Tone
When designing an app focused around using AI to lead interaction, it is important that users feel:
-
Comfortable being vulnerable
-
Heard
-
Supported
-
Understood
-
Like I don’t have to pretend that I am okay
Who make the interaction feel more natural and human, the following key words were used to train the AI how to talk to users.
Key Words: Empathetic, Understanding (of chronic pain), Supportive, Encouraging, Conversational, intelligent but not condescending, Warm, Calm, Clear, Thoughtful, Personable
To get to this mutual understanding of voice and tone cross-functionally, we had a voice and tone workshop to help everyone understand the importance how you talk to someone not just what you say.
-
Who are our users? Develop a full understand of our users and their daily struggles. Understand their "why" and how other products may have failed them-- how this product can be different.
-
Empathy Mapping Be able to see our product through the lens of the user.
-
Voice & Tone level setting Why does design feel this is essential to the product?
-
Establish Critical User Journeys What do our users want and need out of this product-- value props
-
AI Persona Who are some character and people you would want talking to you? What characteristics do they share? How might we incorporate these traits into the AI
-
Consolidate findings
Conversational AI Integration
How might we leverage AI to help make this product feel and act personalized to each user?
Leveraging an LLM (large language model) , we are able to upload relevant data and content to the AI so it can act as a resource and even recommend content based on each users pain profile.
In the "next-gen" design of Fern, we've integrated AI intents, AI determinations, and criteria-based recommendations to create a more personalized and intuitive experience for users managing their pain. Here's how each component plays a role in the design:
1. AI Intents
-
Purpose: Understanding user goals (what users want to do) and actions (what they report they actually did), Fern’s AI analyzes inputs from user interactions (e.g., pain levels, activity tracking, and mood) to infer their intentions.
-
Application: The AI detects when a user intends to seek pain relief, relaxation, or mindfulness exercises based on their behaviors and patterns. For example, if the user logs higher pain levels after certain activities, Fern may prompt them with relevant exercises or coping mechanisms.
2. AI Determinations
-
Purpose: Making autonomous decisions based on data analysis, the AI can determine the next steps or content to suggest.
-
Application: When the AI identifies trends in a user’s pain progression, what the user finds the most motivating or encouraging, relevancies between their why and quick win suggestions, daily check ins (mood, pain, goal completion), it can autonomously suggest changes in the user’s care plan. For example, the system might recognize when a user’s pain is consistently high during a certain period of the day and recommend adjustments, like trying a new mindfulness routine, stretches, therapeutic exercises or a different goal-setting approach.
3. Criteria-Based Recommendations
-
Purpose: Using specific criteria, the AI provides personalized recommendations for interventions based on a set of rules and data inputs.
-
Data Inputs
-
PROMIS assessment
-
Pain Areas
-
Weekly PEG pain interference
-
Users Why
-
Goals
-
Mood
-
Bio metrics
-
Moving min
-
Sleep quality + length
-
-
-
-
Application: Fern’s AI uses criteria like pain area, user goals, activity level, and past success rates to recommend tailored exercises or coping mechanisms. For example, if a user is primarily dealing with chronic back pain and responds well to movement, the AI might recommend yoga sessions or physical therapy techniques focused on mobility.
Through these design elements, Fern enhances user engagement, supports decision-making, and drives more effective pain management tailored to each individual’s needs.
Onboarding & Intake
One of the main issues we needed to address in this redesign was user intake and onboarding.
The main opportunities:
-
Reduce time to finish
-
Move background data to profile to be filled out asynchronously
-
Utilize conversational AI
-
Expectation Setting
Completing intake can feel like a task or chore and it is our job to figure out how to make it more engaging and leave something to be desired in this program. Deriving from a clinical requirements workflow, we made the decision to reduce time to finish by only asking necessary questions and moving additional background info to be collected in the user profile once they are in the program. This drastically cut down on the amount of information user's needed to fill out. From there, we utilized the conversational element to ask questions to make it more approachable and friendly.
We tested our new intake flow with chronic pain users and the results were clear... People need Fern. There was an overwhelming sense of wonder and admiration for the product we are developing and every single user remained engaged and wanted to continue in the product. This conversational approach "felt different" and allowed for further personalization with the generative AI able to take what the user said and validate their feelings and data.
Emotionally Aware Personalization
Fern AI creates a "pain profile" for each individual user which means it stores specific information for each user. From this, we are able to dynamically pull content based on the user's wants and needs in that moment. Each interaction with Fern the AI becomes more accustomed to you and your needs.
True personalization is about understanding the whole person—who they are, what they need, what motivates them, and how they experience the world. This is where AI becomes a game changer, allowing us to blend clinical insights with a truly human-centered experience. Fern AI doesn’t just provide recommendations; it’s dynamic, intuitive, and constantly learning from the user’s inputs.
Imagine an AI-driven tool that not only tracks physical symptoms but integrates emotional patterns, relationships, and lifestyle choices into its core. We’re talking about a whole health experience that sees beyond the data points and connects with the user’s daily life in real, meaningful ways.
For instance, Fern’s AI doesn’t just ask users to report their pain levels—it looks at movement data, sleep quality, and even their relationships to provide a more comprehensive picture of how chronic pain impacts their overall well-being. If the AI detects that a user’s sleep has been erratic for several nights, it might suggest modifying their activity goals, emphasizing recovery and mindfulness. This is AI that doesn’t just offer generic advice; it’s intelligent enough to modify its approach based on real-time data, creating a fully adaptive experience.
Goal setting
Previously, goal setting was something you talked about with your human health coach, this means it was not dynamic and not integrated into the program.
Main opportunities:
-
Make goal setting a core feature in the program
-
Design an interactive activity to set goals
Progress & Insights
User's are able to see all of their progress at a glance, clicking into each section to view more details if desired. We went with a widget approach to hold the most amount of data, but delivering it in a digestible way. Promoting positive engagement first, users see a "time spent" graph representing their time this week in the program. Metrics like these allow for curiosity and positivity to bring them back in to see their insights.
Fern’s AI continuously monitors progress using real-time data from wearables and user-reported outcomes, like PEG pain assessments and personal goal-setting inputs. When a user hits a rough patch or is slow to reach a milestone, the AI recalibrates the plan. Rather than suggesting static, one-size-fits-all solutions, it intelligently offers modified activities or routines that fit the user’s current state. This dynamic goal tracking ensures that the experience remains fresh and aligned with the user’s ongoing journey, reducing the drop-off often seen in long-term health app usage.
THE PROCESS
User Tester Quotes
"It looks like this app was built by someone who understands chronic pain."
bottom of page