Designing Trustworthy AI for Mental Wellness
Through research, iterative design, and collaboration with stakeholders, we refined the Ideal Customer Profile (ICP), converted user insights into design ideas and delivered a product validated by users and recognized by leading investors in Silicon Valley.
Specific designs and product details cannot be shared due to NDA restrictions.
About Aura
Aura Health is a leading AI-driven mental wellness platform with over 7 million users worldwide. Launched in 2017, it offers personalized audio content such as guided meditations, sleep stories, and CBT sessions. Using AI, Aura tailors recommendations to individual needs: fostering emotional well-being, improving sleep and stress management. Backed by prominent investors and led by Forbes 30 Under 30 founders, Aura’s mission is to make mental wellness accessible and engaging.
My Role
As Lead UX Designer, I led the new AI product from the ground up, working with founders in Silicon Valley to define the vision and approach. I quickly learned about mental health, identified user needs, and refined the ideal customer profile (ICP). Through fast iterations and close teamwork with engineering and data science, we solutions based on user feedback. Resulting in optimal information architecture, intuitive wireframes and designs recognized by leading investors and founders, driving the product toward market fit.
The Challenge
The project presented three main challenges:
🚩 1. Finding the right Ideal Customer Profile (ICP):
Identifying the most engaged and receptive audience through user research, which involved analyzing feedback, recognizing patterns, and reshaping our assumptions to align with the needs of this newly discovered user group.
🚩 2. Designing AI that people trust
We noticed that many users were finding difficulties in using AI as a tool to get guidance and manage their emotions. The challenge was ensuring that the product was designed in a way that users find it safe and reliable to use
🚩 3. Delivering designs that feel non-judgmental
Through user research, we found that it was important to create a product that gives users a safe space so that they feel encouraged to use it. It was crucial for our designs to empathize with users without making them feel assessed.
Process & Approach
User Research & Defining the Ideal Customer Profile (ICP)
By gathering insights from a diverse group of users, we recognized patterns that led us to discover a new, highly engaged audience. We tracked NPS scores for interview participants and monitored patterns in the demographics of high scorers. Qualitative responses helped us understand user motivation and context, revealing that most promoters shared similar traits and contexts. We refined our assumptions and adjusted our focus to meet the needs of this group, aligning the product more closely with their expectations and preferences. This iterative process of listening, analyzing, and adapting was key to shaping a product that resonated deeply with the right users.
Ethical AI & Trust-Centric Design
I collected user feedback and iteratively developed innovative ideas to incorporate these insights while validating with users and stakeholders to ensure the design met both user and business needs. The design was aimed to build trust and encourage use. We emphasized on user safety during the onboarding and developed new features that give users control over data collection and personalization. We crafted the user experience to feel supportive and personalized rather than judgmental, ensuring it resonated with users' emotional needs.
Designing a supportive user experience
Real-world feedback played a key role in shaping these design choices, helping us refine the interface to better address user needs. With each iteration, we reassessed insights from our Ideal Customer Profile (ICP), ensuring that design decisions aligned with the expectations and preferences of our target user segments. This process of continuous feedback and adaptation allowed us to create an experience that felt both intuitive and emotionally supportive.
We introduced data visualizations and thoughtful design elements that created a non-judgmental and supportive user experience. For example:
- We came up with abstract visualisations for emotionally sensitive information such as emotion and mood data.
- We routed extreme negative edge cases towards professional help
- We balanced negative and positive information so that users don't feel intimidated by the design
Collaborating with the founders and cross-functional teams
Nothing great is built alone. I worked closely with Aura’s leadership to define the product vision and align it with both user needs and business goals. I presented design insights to stakeholders, integrating their feedback to improve scalability. I worked with technology and data science to understand the technology and give optimized solutions.
Impact & Key Learnings
✨ 1. Prioritising Ethical Design
AI for mental wellbeing has to be designed responsibly ensuring there is no misinformation, biases, or unexpected harm. Products should help humans rather than preying on their vulnerabilities.
✨ 2. Building trust is super important
Users need to feel safe and supported, knowing that the technology is designed with their best interests at heart.
✨ 3. AI’s potential in mental wellness is just beginning
We are only scratching the surface of what AI can do to help people build healthier habits, manage stress, and feel supported in their journeys.
The final design was not only loved by interview participants but also validated and appreciated by prominent investors and founders from Silicon Valley, reinforcing its potential.
