Darshi Shah

Enhancing Hinge Health's Exercise Recommendations with AI-Driven Menstrual Health Integration

I led the design and strategic direction of this project, focusing on integrating an AI-driven solution to enhance exercise recommendations for individuals who menstruate. Collaborating closely with industry partners at Hinge Health, including a Product Designer, UX Researcher, and Product Manager, I worked to align the design vision with both user needs and business priorities around menstrual health integration. Throughout the process, I sought feedback from healthcare professionals, product design experts, and end users to ensure the solution was clinically relevant, user-centered, and technically feasible.

Role
Lead Product Designer, UX Researcher
Timeline
10 months (Aug '24 – May '25)
Core Responsibilities
Product Design, Product Strategy, Research, Visual Design, IXD
Winner · 1 of 2 from 74 teams
Nominated · iF Design Student Award
Nominated · UX Design Awards, Healthcare
Hinge Health menstrual health integration — case study banner
Overview
Challenge

Strategic Expansion into Women's Health to Elevate NPS and Drive User Growth at Hinge Health

Hinge Health aimed to expand into the women's health space to reach new user segments and improve overall satisfaction, with a parallel goal of integrating AI into their platform as a cost-effective enhancement. With broad possibilities, ranging from new menstruators to pregnant and menopausal users, the challenge was to identify a focused, high-impact entry point. This required aligning user needs with business goals, and iteratively developing and testing a solution to evaluate its feasibility, usefulness, and long-term value.

Solution

Smart AI-powered exercise recommendations tailored to your menstrual cycle for personalized pain relief

We designed an AI-powered, cycle-aware exercise recommendation system within the Hinge Health app. By factoring in each user's menstrual cycle, symptoms, and energy levels, the AI delivers personalized, evidence-based exercise plans that adapt in real time. This ensures users not only manage pain more effectively but also stay active and feel supported throughout their cycle without disrupting existing routines.

New to the appMenstrual Insights

The core of our solution is the new AI-powered Menstrual Insights section. Here, users can:

  • Understand why specific exercises are recommended for their current cycle phase, with AI providing explanations and adjustments
  • Log their periods and study cycle patterns, with AI surfacing trends over time
  • Access trustworthy resources about menstrual health

This intelligent combination of context, tracking, and education helps users feel informed, supported, and in control.

RedesignedAdaptive Exercise Flow

We redesigned the daily workout flow to be adaptive and flexible, using AI.

  • Users preview exercises and required equipment for the day
  • Pain levels can be adjusted in real time, with AI dynamically modifying intensity and exercise recommendations

This AI-driven adaptability empowers users to stay active even on lower-energy days, while keeping workouts safe and manageable.

RedesignedProgress

Feedback revealed dissatisfaction with the old progress view in Hinge Health. Users wanted richer, more meaningful insights. We redesigned the section to show exercise progress over time, including weekly, monthly, and yearly views. The new design highlights trends that motivate consistency and reinforce healthy habits.

ExpandedOnboarding

We expanded the onboarding flow to capture users' cycle information and patterns. To minimize extra effort, we also enabled Apple Health integration, so users could sync their existing period data without logging it separately in multiple apps.

The pivot

The Decision That Shaped Everything: The user does not trust a chatbot with its health

The original concept included a visible AI chatbot for menstrual health questions. It felt like the obvious move: AI product, conversational interface, instant answers.

Concept testing said otherwise. 13 of 19 participants were skeptical of AI for health information. They worried about misinformation, and they didn't want to negotiate with a bot about their own bodies.

What we planned

A visible health chatbot

Conversational answers to menstrual health questions, front and center in the app. Cut after concept testing.

What I built instead

AI in the backend

It still personalizes everything: intensity, focus, and recommendations shift with cycle phase. But users never talk to it. They see a plan that fits how they feel today, and a plain-language reason for each exercise. Personalization without the trust tax.

This one decision drove the rest of the design. Every feature after it had to pass the same test: does this help without asking users to trust a machine with their health?

Additional info
Discovery

Discovering the Need to Address Menstrual Pain

Research approach · 5 methods
1
Survey
40+ responses
2
Interviews
5 participants
3
Concept testing
19 participants
4
Expert walkthroughs
4 specialists
5
Usability testing
4 participants

To better understand our users and the impact of menstrual pain on their daily lives, we conducted a targeted survey.

Survey

40+ responses from individuals aged 18–30 experiencing menstrual pain and symptoms of Primary Dysmenorrhea.

Interviews

5 participants sharing in-depth personal experiences.

Goals:

  • Identify the specific challenges users face during their menstrual cycle
  • Understand existing coping strategies for managing pain
  • Explore the role of exercise in participants' lives and how it is affected by menstruation

Key Insights:

1

Tracking tools lack actionable value

While many users rely on cycle tracking apps before and during their period, they often find the insights too generic and not directly useful for managing pain or activity.

2

Menstrual pain disrupts routines

Pain frequently forces users to reschedule or cancel planned activities, including exercise, work, and social engagements, impacting both productivity and lifestyle consistency.

3

Emotional well-being is heavily impacted

Users experience severe mood changes before their period and significant physical discomfort during it, leading to reduced quality of life and emotional resilience.

4

Strong desire for lifestyle changes, especially exercise

Many users see exercise and other healthy habits as key to managing menstrual symptoms, but pain and low energy during their cycle often limit their ability to stay consistent.

5

Lack of trustworthy resources

Users struggle to find reliable, evidence-based information about their bodies and menstrual health, leading to uncertainty in managing symptoms effectively.

Opportunities

Creating the Foundation for Concept Development

Insights from the survey and user interviews not only revealed key pain points and needs but also helped us clearly define our target persona. By mapping recurring patterns in experiences, goals, and challenges, we built a representative user profile to guide our design decisions. This persona ensured that every requirement and concept we explored addressed the needs of a clearly identified user, rather than a broad, generic audience.

Target user persona built from research patterns

Building on the insights from our survey and interviews, we translated user needs into clear design requirements that would guide our solution. These requirements ensured our concepts stayed grounded in real user challenges while aligning with project goals.

Design requirements translated from user needs

We then explored the opportunity space through collaborative brainstorming and competitor analysis. This allowed us to identify potential features, understand market gaps, and gather inspiration from existing solutions, laying the groundwork for informed and targeted concept development.

Opportunity space from brainstorming and competitor analysis
Concept validation

Testing Early Concepts with Target Users & Experts

Following our brainstorming, we created low-fidelity sketches of our proposed features and tested them with 19 participants (13 users and 6 experts) representing our target user group.

Low-fidelity concept sketches used in testing

Objectives:

  • Gather opinions on the value and usability of an AI chatbot for health support
  • Evaluate the layout and concept of the activity tracking section
  • Evaluate the layout and concept of the menstrual information section

Findings:

1

Prefer plain, scannable language

Users prefer less scientific, straightforward language when viewing period information, making sure it is easily scannable.

15 / 19 participants
2

Want the reasoning behind recommendations

Users were curious about the reasons behind specific exercise recommendations and wanted to understand them better to learn more about their bodies.

14 / 19 participants
3

Skeptical of AI for health information

Users were skeptical about relying on AI for health-related information and felt that a chatbot wouldn't add value to the program.

13 / 19 participants
4

Prefer a visual display of activity

Users preferred a more visual display of their daily activity, using elements like rings and color cues that align with current industry standards.

14 / 19 participants
Mid-Fi feedback

Refining Through Expert Walkthroughs

Based on our initial concept testing, we made several design refinements:

  1. Moved the Focus section to the home page and added contextual explanations
  2. Created a dedicated Menstrual section
  3. Added more highlights and trends for better progress visibility
Mid-fidelity refinements to the app sections

We then conducted expert walkthroughs with four specialists in product design and women's health.

Expert walkthrough sessions with specialists

Key Feedback:

  • Repetition in playlist information made the experience feel redundant.
  • Annual activity trends would provide a broader perspective on progress.
  • The language tone could be clearer and more accessible to better engage users.

Based on this feedback, we incorporated these changes into our final prototypes.

Final prototype updates from expert feedback
Refined prototype screens
Final user-feedback

Validating the Final Design with Target Users

We conducted high-fidelity, task-based usability testing with four participants from our target user group.

Goal: The goal was to assess the clarity, usability, and perceived value of the integrated menstrual health features within the Hinge Health app.

Results:

  • Participants appreciated learning why certain exercises were recommended and found this information helpful for managing their period.
  • The added personalisation in exercise planning addressed a major barrier to maintaining an exercise routine.
  • Users described the problem space as meaningful and expressed enthusiasm for the solution's potential impact.
  • SUS responses indicated strong agreement that the system was easy to use and something they would like to use frequently.
Reflections

Lessons Learned and Directions for Growth

This project highlighted the value of agility and resilience when navigating evolving requirements and user feedback. Clear communication within the team proved essential for aligning on priorities, and timeboxing helped us maintain momentum while meeting deadlines.

Looking ahead, opportunities remain to enhance the solution, such as integrating mood tracking, expanding exercise options tailored to menstrual phases, and conducting further user testing to uncover new insights and refine menstrual health features.

Reflections and future directions for the project