How to Redesign Retention in Mental Wellness Apps: The Zerenly Case

When we analyzed the metrics for Zerenly, a mobile app designed to transform the way people manage their emotional and mental well-being, we noticed a recurring pattern: users would arrive, log how they felt, explore an AI-generated analysis—and then not return as frequently as expected.

The emotional logging feature worked well to attract new users—primarily women, who make up 80% of the user base—but something was missing to generate the sustained engagement we were aiming for.

This led us to a key question: how can we improve user retention in an emotional wellness app? The challenge wasn’t just to increase daily engagement, but to make the Zerenly experience leave a deeper and more meaningful impression.

‍

That’s when it clicked: we had to redesign from the purpose

This diagnosis led us to a crucial realization: we couldn’t think about retention without first creating a real emotional connection. We moved from a short, linear journey—emotion logging followed by an analysis—to a richer, more meaningful process.

‍

Redesigning the flow: from functionality to purpose

The new flow introduces a key step that changes everything: the follow-up question. A brief, empathetic, personalized prompt designed to deepen the experience and leave a lasting impression.

This change goes beyond a functional tweak—it’s a deliberate choice aligned with Zerenly’s purpose: helping people better understand what they feel, and not stopping at a superficial log.

After selecting the emotion, sub-emotion, and describing the experience, the app now offers a personalized question that invites reflection. It’s not a question meant to be answered immediately—its goal is to resonate, support, and motivate users to return.

‍

The power of personalized follow-up questions

The follow-up is inspired by the principles of ontological coaching, which emphasize the power of thoughtful questions as a trigger for self-awareness. Each question is written with intention: it aims to validate the experience, provide warmth, and spark curiosity.

Some examples of questions generated during our internal testing:

• “What do you think might be behind the discomfort you're feeling?”
• “What would you like to change about your work environment to feel more comfortable and motivated in the morning?”

These questions are made possible by an AI model that receives contextual information (selected emotion, user description, age, gender, and other relevant data). From a carefully crafted prompt, the AI generates an empathetic, brief, and open-ended follow-up. In fact, the prompt is one of the most strategic components of the entire flow, as it clearly defines the direction of the interaction.

‍

How to Integrate Artificial Intelligence with Empathy

Designing this experience posed a significant technical challenge. Inserting an intermediate step into an already functioning flow meant answering several questions:

• Where is the follow-up question generated?
• When is it displayed?
• What happens if the AI fails?
• How are the data integrated with emotional insights?

To address these, we created a new endpoint that receives the user's emotional log and returns the generated question. Capturing and displaying this information cleanly was key to preserving the user experience.

The core solution was designing an “emotional middleware” that kept the architecture organized, testable, and open to iteration. This allowed the follow-up question to be seamlessly integrated into the flow, without friction or unnatural pauses.

We also built a system that stores valuable data at every step: emotion, sub-emotion, description, the generated question, and the user’s response, if any. This enriched dataset allows us to better understand real-world usage.

‍

At the Heart of It All: A Well-Crafted Prompt

A central part of the success of this feature lies in the prompt that drives our AI model. Crafting it required deep awareness of the goal: if the prompt isn’t clear and precise, the experience suffers.

That’s why we wrote a detailed prompt that defines exactly how the AI should behave. Here are a few excerpts from that prompt, which we consider a strategic piece of the project:

‍

This prompt ensures that each question aligns with the brand’s tone, conveys emotional support, and fulfills Zerenly’s mission: to accompany without intruding.

‍

The Role of Microinteractions in Boosting Retention

Beyond functionality, we wanted the follow-up to feel warm and personal. That’s why we introduced microinteractions: a loading state with a shifting gradient, subtle glow effects to suggest something is happening behind the scenes, and small visual gestures to reinforce the emotional connection.

These microinteractions were created in Jitter and exported as Lottie animations—lightweight and ideal for seamless integration.

‍

It’s Not All About AI: Language as an Emotional Bridge

Zerenly has always stood out for its empathetic, approachable, and genuine tone. Every word is carefully chosen, recognizing that users may be experiencing vulnerable moments.

For this new flow, we organized a team-wide workshop to explore different ways of naming, questioning, and supporting. Each participant brought both professional and personal perspectives, helping us strike the right balance between clarity, compassion, and purpose.

As a result, we created a language that encourages reflection without pressure, provides support without overprotection, and respects the emotional process.

‍

Results and What’s Next

Beyond the metrics, this process allowed us to reconnect with Zerenly’s core purpose. We know that retention isn’t just about screen time—it’s about the resonance we leave with those who choose our client’s app.

Today, we’re convinced that a well-crafted question can spark internal change. And if someone decides to return because they felt supported, then all of this has been worth it.

At Paisanos, the journey continues: we’re launching A/B tests to understand how this feature impacts emotional insights, user relationships with the app, and real retention. But beyond any data point, we believe that technology can be an ally in creating more human experiences.

Because in the end, it’s not just about logging an emotion—it’s about feeling that someone (or something) understands you, hears you, and leaves you thinking.

Get the full story behind how we redesigned retention for Zerenly here.

This content was created by Josefina Staudenmaier, Santiago Crippa, and Lucas Gallucci, part of the team dedicated to designing and improving the Zerenly experience, united by the belief that technology can be a bridge to a more genuine connection with how we feel.

‍