Bea Better Eating: Building from Scratch Case Study

From Zero to Launch: designing an AI-powered emotional-eating therapy app built to scale BEFORE OpenAI was available for public use.

Bea Better Eating was created to move beyond diet culture — offering compassionate, AI-powered support for emotional eating.

As product lead, I combined design thinking with clinical hypnosis and Ericksonian methodologies to build a scalable, empathetic platform that helps users understand why they eat, not just what they eat.

This case study captures Bea’s earliest stage—from foundational research through concept definition, prototyping, and validation.

PHASE 1

Research: Discovery & Problem Definition

In collaboration with Dr. Sera Lavelle, I conducted empathy interviews with clinicians and individuals experiencing emotional or binge eating. Her de-identified clinical insights were invaluable, helping us identify core behavioral patterns and personality traits common in disordered eating — insights that shaped both our tone and product direction.

  1. Diet apps felt judgmental and rigid.

  2. Users were exhausted by willpower-based advice.

  3. They wanted compassion, not calories.

Insights

Build a non-judgmental space that addresses emotional drivers behind eating, grounded in hypnosis and behavioral science. An accompanying Insight Dashboard will give people a better idea of trends and triggers during their lows and highs.

Core Opportunity

PHASE 2

Define: Concept & Strategy

We designed Bea as a supportive companion, not a tracker—combining AI guidance, emotional safety, and small, achievable habits.

  • Therapeutic experiences: Grounded in Ericksonian methodology and evidence-based hypnotherapy techniques proven effective in addressing emotional eating.

  • Language pacing and tone: Inspired by hypnotic cadence and positive suggestion to help users access self-awareness without judgment.

  • Root-cause focus: Emotional triggers over calorie counting.

  • Safety: Trigger-free UX and filtered language to prevent harm or shame-based interactions.

  • Scalability: Modular feature architecture and an empathetic, consistent brand voice.

Key Pillars:

Every design and product decision was made with both clinical integrity and scalability in mind — ensuring Bea could maintain psychological safety while growing to serve diverse audiences.

PHASE 3

Design & Prototyping

With our core strategy defined, I translated Bea’s therapeutic principles into a digital experience that felt safe, empathetic, and scalable.

  • AI chat interface: Simulated empathetic, real-time conversations guiding users through Knowledge, Motivation, and Meditation activities — modeled on hypnotic pacing and Ericksonian dialogue.

  • Eating Insight dashboard: Explored how structured reflection could surface emotional triggers and patterns over time.

  • AI mind maps: Defined Bea’s personality traits, language tone, and conversational flow to ensure consistency across interactions.

  • Usability testing: Validated tone, inclusivity, and overall comfort with early users drawn from research participants.

  • Also tested Eating Insight Dashboard and extended Bea chat flows to assess emotional trust, pacing, and clinical scalability.

Focus Areas

PHASE 4

Market Validation: Results & Learnings

After multiple rounds of prototype testing, we synthesized findings and prepared Bea for full development.

  • Validated therapeutic model: Hypnosis and Ericksonian tone resonated strongly, building trust quickly.

  • User feedback: Participants described Bea as “comforting,” “understanding,” and “non-judgmental.”

  • Interface clarity: Simplifying flows and language improved users’ sense of safety and reduced overwhelm.

  • Technical foundation: Established analytics pipeline and AI architecture for longitudinal tracking in the next phase.

  • Next step: Begin controlled rollout and post-launch data monitoring to for potential engagement and retention refinements.

Design & Validation Insights

Begin controlled rollout and post-launch data monitoring to for potential engagement and retention refinements.

Next Steps:

Bea’s early design phase proved that clinical hypnosis can scale digitally without losing empathy and was absolutely filling a gap in the wellness space. (Note that this phase was completed before OpenAI became available for public use, and agentic-AI was not widely used.)

Through careful testing and iterative design, we transformed a therapeutic framework into an accessible, human-centered product.

This groundwork—psychological, technical, and empathic—set the stage for Bea’s next chapter: nine months of data-driven iteration and retention improvement.

REFLECTION

MORE BEA CASE STUDIES

BACK TO RECENT WORK