Context

Postpartum care often drops off just as support needs increase. New mothers are left to navigate physical recovery, identity change, feeding questions, sleep deprivation, and mental health concerns through fragmented resources and contradictory online advice.

Mothering Together: Designing for AI was a 2023 buildspace s5 concept exploring how evidence-based AI support could meet parents at scale while still respecting the sensitivity of maternal care.

My role

I led UX research, AI interaction design, and high-fidelity prototyping in Figma, with research and testing workflows supported by Optimal and Maze. The work focused on concept validation rather than production machine learning.

Constraints

Because the product sits close to health guidance, the AI experience needed clear disclosure, careful escalation, and obvious limits. The audience was vulnerable, the subject matter was emotionally charged, and the prototype had to communicate trust without overstating capability.

Research

Mothers described a gap between what they needed and what existing tools provided. They could track symptoms or join forums, but expert-backed answers were harder to find. Apps often created more sorting work instead of reducing uncertainty.

The opportunity was not simply adding AI. It was designing an AI support flow that felt warm, explained itself, and knew when to stop. Market signals suggested room for innovation, with maternal care apps projected to reach $1.3B by 2032.

Timeline

• 01 Discover problem framing

• 02 Research interviews + competitive review

• 03 Design Elle AI personality + chat UX

• 04 Ship prototype + Maze testing

Solution decisions

• Elle AI became a warm, empathetic assistant with evidence-based prompts instead of a generic chatbot voice.

• AI onboarding was gradual, introduced after context was set so users understood its role before relying on it.

• Urgent or complex queries surfaced an explicit AI-to-human escalation path instead of burying safety guidance in fine print.

Outcomes

The project validated the main AI chat flows, produced a high-fidelity prototype, and refined prompts around disclosure, tone, and handoff moments. Future paths included community support, telehealth integration, and expert review loops.

What I learned

Maternal health fails when AI is treated as the product rather than the interface. Trust comes from the flow: what the system asks, how it frames confidence, and how clearly it hands off to people when the stakes rise.

Systems diagram

Support flow: Onboarding → Elle AI → Guidance → Triage → Human expert.