Product strategy · research · UX/UI · prototyping
AI career support guide Live beta
HerNext
I turned a broad community idea into a focused AI guide for women navigating career uncertainty: it proposes, she decides.
01 / The brief
Make career uncertainty feel workable.
HerNext helps women turn private reflection into possible career paths, relevant stories, and meaningful connections.
I conceived, researched, designed, and built the product. Rapid coded prototypes helped me turn an expansive community vision into a focused, responsible AI experience.
Three product iterations from first concept to live beta.
Started in Claude Code, designed and prototyped in Claude Design, then built and shipped with Codex. GitHub and Vercel supported versioning and deployment.
A working mobile web product for invited users.
Target audience
Women navigating career uncertainty.
HerNext meets women at different points in change: before the path is clear, while they are building, and when sharing their experience can help someone else move forward.
Exploring
Finding language for a possible next chapter.
Building
Turning an idea, project, or independent path into motion.
Connecting
Learning from and offering support to women on a similar path.
Career reinvention spans inspiration, reflection, exploration, and connection, but existing products split those needs across disconnected experiences.
I used feedback and live prototypes to narrow the concept, then designed an agentic workflow that makes its reasoning and approval points visible.
A live AI-guided beta that turns a private signal into a grounded next step while keeping the user informed and in control.
Each cycle made the promise narrower and more useful.
AI-assisted building compressed learning cycles.
Reflect → explore → connect.
The Guide proposes; the user decides.
02 / The inflection point
The community felt warm. It did not yet feel essential.
The first concept offered many ways to participate, but no single essential job. Feedback revealed the sharper opportunity: help women privately make sense of what could come next, then connect them to relevant paths, stories, and people.
The promise was too broad
Stories, wishes, matches, circles, and support all felt valuable, but there was no single reason to return.
Community required momentum
The experience depended on active members before it could become useful to someone arriving alone and uncertain.
The private moment was underserved
The highest-value moment happened before a public post: making sense of a change that was still hard to name.
A community for women building what’s next
- Value grows with participation
- Many parallel behaviors
- Public expression comes early
An AI guide for making sense of what’s next
- Turns a first private reflection into tailored next steps
- One connected behavior loop
- Matches women on similar paths to connect and learn
03 / The product evolution
Three iterations reshaped the product and the screens.
I treated each coded prototype as a question the next round of feedback could answer. That made it possible to test the product model, information architecture, visual tone, and role of AI in context.
- V1 · Community
Make women’s journeys and support visible.
Mapped onboarding, stories, profiles, matches, wishes, and support into a complete mobile concept.
- V2 · Focus
Organize discovery around intent.
Separated exploring from building, simplified overlapping feeds, and turned vague encouragement into concrete support actions.
- V3 · AI guide
Let private reflection power the next step.
Connected reflection to possible paths, personalized discovery, and introductions through an explainable, approval-based workflow.
Feedback: “As a user taking an assessment, I want to know why and how it will impact my app experience.”
I moved the product promise up front, explained how each answer shaped the experience, and organized onboarding around what the user needed rather than a fixed professional identity.
Feedback: “I think you were trying to show Stories that resonate with the user and maybe a list of women who match their profile or needs? That could be more useful than showing who are actively buildings vs. who are exploring/thinking of building.”
I removed competing feed models and separated the two clearest jobs: Stories for inspiration and Matches for relevance.
Feedback: “I feel a sense of pressure to perform since all the examples are about women who’s starting/building something. I would be mindful of that - if that’s your audience, that’s fine. But not all women will be in that stage or mindset. I would consider something else more catering to the different stages.”
Wishes gave uncertainty a place before it hardened into a goal. That insight became the bridge to private AI-guided reflection.
A personalized home connects possible paths with relevant stories and affirmation. From there, women can understand why a match matters, learn from her journey, and connect with context.
04 / The agentic model
An AI Guide that works agentically to turn uncertainty into possibility.
A chatbot would have made AI obvious, but not necessarily useful. I designed the Guide as a quiet layer across the journey: it gathers context, interprets patterns, communicates when confidence is still taking shape, explains its reasoning, and previews a proposed next step before the user approves or rejects it.
Private context
Personal reflections inform the Guide without becoming public content.
Transparent reasoning
Suggestions explain which signals were connected, why, and when the pattern is still taking shape.
Permission before action
The Guide proposes and previews; the user can correct it and must approve before anything moves forward.
The core loop
Connect the dots. Keep her in control.
A user can pause on a story, reflect privately, see how the Guide understood her words, see when its confidence is still forming, and decide whether that thought should shape what appears next.
Private signal→Explained suggestion→User approval
05 / The experience
One connected journey from uncertainty to action.
The final experience does not ask users to perform certainty. It gives them a progression: reflect in private, inspect possibilities, discover useful proof, and connect when the context feels right.
Explore possible paths
Make possibilities concrete without prescribing the answer.
Four focused questions turn ambiguity into language the Guide can build from. The output is framed as a set of hypotheses to inspect and try, not an answer generated on the user’s behalf.
Build a reason to return
Direction, proof, and permission for different emotional seasons.
Sometimes a user needs a path. Sometimes she needs to see how another woman made a change. Sometimes she simply needs encouragement that meets the moment.
From resonance to connection
Meet the right woman.
Private reflection helps HerNext understand the season a user is in. The product explains why a woman is relevant, connects the shared context, and previews a drafted hello before any outreach. If the AI gets it wrong, the user can edit it, send it, or ignore the suggestion entirely.
Come as you are. Move at your pace.
Find stories, possibilities, and language for what might come next.
Make the next step visible and invite the support that creates momentum.
Explore change while staying employed and growing in a current role.
06 / The outcome
A working AI product, not a polished thought experiment.
In about four weeks, HerNext moved from a broad community concept to a live beta with a clear job, a responsible AI interaction model, and a focused loop the product can now test with real users.
Explore the live betaShipped in beta
Prove relevance and trust
- Intent-led onboarding
- Stories and matched women
- Private reflection
- Possible career paths
- Transparent AI reasoning
- Approval before action
Next experiments
Earn the next layer
- Small accountability circles
- Weekly AI digest
- Richer support requests
- Longitudinal reflection
- Direct payment safeguards
- Broader AI coaching