We joined Smoodai in the early stages of their research and experimentation phase, with one shared goal: to bring behavioural depth to a tool designed for emotional support. This meant looking beyond what AI can generate, and focusing on what users need in emotionally complex moments.

In the mental health tech space, how something is said is often just as important as what is said. Many digital check-ins begin with “How are you feeling?”—a well-intentioned prompt, but one that can feel flat or even intrusive if not followed by a meaningful interaction.

We began by guiding Smoodai through a reframing of their early prompt structure. Instead of directive advice or generic sympathy, prompts were restructured to offer light-touch agency and scaffolding. For example:

• Before: “How are you feeling today?”

• After: “Rough day? Want to reflect, distract, or shift focus?”

Another example replaced flat encouragement with reflective nudging:
“You’ve checked in three days in a row—most people don’t. That’s real resilience.”

Small tweaks, big shifts. These changes drew from motivational interviewing, cognitive behavioural therapy, and social norms theory—ensuring that every interaction was behaviourally informed, emotionally grounded, and tailored to the user’s psychological state.

The Strategy: Simple, Behaviourally Rich Foundations

Smoodai didn’t need a complete product overhaul. What they needed was a foundational behavioural strategy that could grow with them.

We began by using the COM-B model (Capability, Opportunity, Motivation – Behaviour), a behavioural science tool for diagnosing barriers to action. Together, we mapped key behavioural moments across the user journey—from first interaction to repeat check-ins.

Here’s how COM-B shaped early design decisions:

• Capability: Could users figure out how to use the app in moments of stress?
  We recommended visual walkthroughs, “micro-practice” flows, and a minimalist interface that doesn’t overwhelm.

• Opportunity: Were users in a context where they felt safe to engage?
  We suggested aligning push notifications with private, quieter moments (like post-school evenings or winding down at night), rather than mornings or mid-day.

• Motivation: Did users see themselves in the experience?
  Prompts began using identity-based nudges: “Many students feel this way post-exams” or “You’re not the only one who struggles with focus during long breaks.”

Alongside this journey mapping, we created a lightweight research kit that Smoodai could use to begin testing:

• Early-stage prompt A/B testing (supportive vs directive tone)

• Emotion sliders at the end of each check-in (to capture change in feeling)

• Drop-off observation prompts to understand moments of disengagement

Importantly, all of these were built with ethical UX principles in mind—ensuring Smoodai was not just measuring engagement, but protecting emotional safety along the way.

We also trained the team on the difference between passive usage and active connection. While many apps measure “time spent,” we encouraged them to focus on “emotional stickiness”—that is, whether users feel better, seen, or motivated after a short interaction.

The Results: From App Prototype to Emotional Ally

In just six weeks, Smoodai moved from a promising AI build to a behaviourally grounded MVP.

Key outcomes included:

• 42% increase in prompt engagement during early A/B tests

• 28% reduction in drop-off during the onboarding journey

• Early signs of positive emotional shift in short-session feedback loops

• A prompt library built on psychological insight, not just AI generation

• A team now equipped to run ethical, light-touch experiments with confidence

More importantly, users in early testing shared that the experience felt gentle, human, and “not like being coached by a bot.”

Smoodai now has a behavioural foundation that can scale with them—and a clear point of difference in a crowded mental health app market.