Mothering Together

Designing Elle for trust

Mothering Together is a postpartum support platform I’ve been building since November 2022. It started as UX research and a prototype (including a Designlab UX capstone), became an AI-feature prototype through Buildspace in 2024, and launched as a full MVP on April 19, 2026 — 19 days from first commit to launch.

The live product addresses what happens between appointments — the 3 AM question, the daily check-in, the moment when something feels wrong and there’s no one to ask. Elle is an AI companion for personalized, evidence-based postpartum guidance: conversational, warm, grounded in clinical evidence. Not a chatbot.

Clinical advisors: an emergency medicine physician, a nurse practitioner and a retired psychiatric nurse. (My partner, my sister, and my mother.)

I designed the conversation UX, the prompt architecture, and the guardrails that keep Elle within evidence-based guidance. The AI layer is powered by Claude Sonnet 4.6. Elle includes a crisis and safety module and “is this normal?” coverage starting with physical recovery.

One of the more deliberate design decisions was the content labeling system: responses are visibly tagged as either “AI guidance” or “expert reviewed.” A postpartum mother asking about symptoms she’s frightened by needs to know whether she’s reading a model’s interpretation or a clinically reviewed answer. The labeling isn’t decorative — it’s the product’s trust mechanism made visible. Elle also always shows a crisis resource button in the top corner, so no matter what a user is talking about, one tap gets them to help.

This traces back to early research: mothers reach for friends, family, and social media because those sources are faster and warmer than clinical resources — but they often provide misleading information. The question was never how to make content more clinical. It was how to make credible information feel as accessible as a knowledgeable friend.

The deeper design challenge was ElleMemory — the system that lets Elle store and retrieve longitudinal context across a mother’s postpartum year. A mother who trusts Elle through her first year creates a record of recovery, mood, sleep, and feeding that no other product owns. This isn’t just personalization. It’s the foundation for a product that gets better over time because it actually knows you.

The problem

Maternal mortality rates in the U.S. have been rising for decades. 60% of pregnancy-related deaths are considered preventable. Between 10–40% of mothers don’t attend their postpartum follow-up. Organizations like ACOG, AHRQ, and the Commonwealth Fund point to the same structural failure: fragmented care, a single 6-week appointment, and mothers left to figure out the rest on their own.

Existing products don’t solve this. Lifecycle apps like Ovia, Flo, and Glow treat postpartum as the third trimester of a pregnancy app — the mother’s recovery sits next to baby sleep guides. Postpartum-specific apps focus on tracking, which helps record symptoms but not answer a 2 AM question. Employer platforms like Maven reach only mothers with that benefit, and mostly for clinical visits — not the daily “is this normal?”

I evaluated five postpartum-adjacent resources in depth — MedlinePlus, March of Dimes, What to Expect, AAFP, and Postpartum Support International. Gaps were consistent: none covered the full 12 months post-delivery, none offered personalization, none gave mothers direct provider access outside a scheduled visit.

Nobody had built something that centers the mother, not the baby — and gives her trusted, evidence-based guidance in real time.

Foundation research

Before Elle, I ran interviews, surveys, and a card sort to learn how mothers look for postpartum information — work that began as a Designlab UX capstone and still anchors the product.

Mothers wanted mobile-first, text-based answers on the device already in their hand. They were already turning to friends, family, and social media because formal resources felt distant — so credibility had to feel as fast and warm as those informal sources. Emotional and mental health support showed up consistently as the most underserved need.

I ran a remote open card sort (Optimal Workshop) with 56 postpartum topic cards. Five participants completed it; 13 abandoned. The abandonment rate was itself a signal: 56 cards is too much cognitive load for a sleep-deprived population. That became a design principle — the product cannot demand effort. Completed sorts clustered topics by urgency and personal concern, not medical specialty, which shaped an urgency-first information architecture.

Early participants were healthcare-adjacent. The live product has since been used by actual postpartum mothers, which continues to inform Elle’s tone and the content areas people reach for first.

Personas

User personas — healthcare-adjacent participants representing different relationships to postpartum information

Affinity map

Affinity mapping — clustering findings across experience, pain points, and needs

Task flows

Task flows: finding an article, scheduling an appointment, accessing a scheduled appointment

Final prototype

Final prototype — mobile-first screens for postpartum support

←

1 / 4 →

Building it

I’m a designer who ships. The MVP was built on Anything.com for speed — frontend, backend, and data layer for the initial launch. I then migrated to Vercel with Supabase, using Cursor agents to build and push to GitHub.

The migration was driven by a real constraint: the product needs HIPAA-aligned infrastructure as it moves toward clinical workflows. Vercel supports BaaS under a shared-responsibility model. Payments run through Stripe with a production webhook.

The design work on top of that stack is where most of the decisions live: prompt architecture, triage flows for crisis escalation, content guardrails, and clinical-source validation. These aren’t engineering decisions. They’re information architecture and conversation UX — the same discipline, applied to a new surface.

Early results

In the first 30 days after launch, with no paid acquisition — just organic posting in postpartum communities and on social (Threads, LinkedIn, Facebook, X):

Roadmap

The roadmap is sequenced around trust. Each layer adds depth only after the previous one has earned credibility with users.

After physical recovery (live now), the next layers are: emotional and mood support, daily reflections, feeding and lactation support, a maternal recovery tracker, a mother-centered postpartum timeline, and a virtual care directory.

Reflection

This is the project where everything I’ve done — library science, information architecture, HRIS, UX research — converges into a single problem worth spending years on.

The thing I keep coming back to: the way you organize information determines whether someone can use it. A clinical database and a postpartum app have different users, different contexts, and different stakes. But the underlying question is the same. Can the right person find what they need, at the moment they need it most?

For Mothering Together, that moment is 3 AM, alone, with one hand free. And the answer has to be yes.

• UX research
• Mobile design
• Info architecture
• Conversation UX
• AI design
• Healthcare
• Founder