"SkillAxis"
Designing a Production-Ready AI EdTech Platform from MVP to Investor Demo
SkillAxis was a pre-seed AI EdTech platform with a real problem — users weren't converting because they didn't trust the product enough to pay upfront. As a contract UX/UI designer, I redesigned the landing page to build credibility, streamlined onboarding flows that were losing people mid-setup, and built 20+ learner-facing dashboard screens on a scalable design system. The pressure was real — a full MVP delivered in 48 hours for an investor demo. My job was to make an early-stage product look and feel production-ready, fast.
Read the Full Case Study Below
Selected screens from the onboarding redesign and learner dashboard.
What is this project about?
End-to-End UI/UX
Designed 20+ screens across learner dashboards, onboarding flows, and landing pages — from initial user research through high-fidelity production-ready screens built on a scalable Figma component library.
Trust-First Design
The platform had a conversion problem — users weren't willing to pay upfront. I redesigned the landing page, and onboarding screens to build credibility through social proof, and real product screenshots, that made the product feel trustworthy.
Delivered Under Pressure
Full MVP interface delivered within 48 hours for an investor demo with real stakes. Used AI-assisted prototyping tools to hit the deadline without cutting corners on quality or consistency.
Project Overview
⏳ Timeline
- September 2025 — January 2026
- Contract Role, Remote
👩🏼💻 Who Was Involved
- Founders/Product Manager, Engineers/ Developers
💻 Remote Collaboration
- Tools Used: Figma, Jira, Slack, Confluence
- Meetings: Weekly Syncs, Daily Standups
- Asynchronous Communication, Shared Docs, Figma Files
🎉 The Result
- MVP delivered in 48 hours for investor demo
- 20+ role-based screens with a scalable design system
- Improved trust signals reduced drop-off on landing page
- Onboarding steps streamlined for faster configuration
What was the main problem?
Conversion & Trust Gap
SkillAxis was asking users to pay upfront — $1 to unlock the full roadmap and a per-session fee to book a coach — but the landing page gave them no reason to believe. No social proof, no real product screenshots, no credibility signals. Combined with an onboarding flow that had too many steps and no progress guidance, users were dropping off before they ever reached the product.
Interface Lacked Polish
The UI had been built by developers without a consistent design system — functional but visually inconsistent and not polished enough for a product asking people to pay or for an investor demo. SkillAxis needed a designer to come in, establish consistency, and bring the interface up to a standard that reflected the quality of the product behind it.
The Research
Before redesigning anything, I conducted 4 structured 1:1 user interviews with early-career professionals actively seeking PM roles — recruited through LinkedIn and personal networks. Each interview followed a structured guide focused on trust, onboarding usability, roadmap clarity, mentorship, and career outcomes.
Participant A — Early-career professional transitioning from a non-PM background, seeking her first PM role in the US
Participant B — Career pivoter with a non-traditional background, exploring PM as a next step
Participant C — Mid-career professional with 4+ years of experience, upskilling toward a senior PM role
Participant D — Experienced professional with 10+ years in strategy, seeking a structured PM career roadmap
01 — Trust before payment
All 4 users asked "why should I trust this?" before committing. They wanted coach profiles, testimonials, and roadmap previews upfront — not after paying.
02 — Confusing self-assessment
The drag-bar for skill level had no reference points. Users didn't know how far to drag. They wanted simple, clear options to choose from.
03 — Personalization gap
Users felt the onboarding was generic. Missing roles like Business Analyst and regions like UK made it feel like the platform wasn't built for them.
04 — Mentorship as differentiator
All four participants cited 1:1 coaching as the primary reason they'd pay — not content. This wasn't a feature, it was the value proposition.
Four Decisions Driven by Research
01
Drag bar replaced with 4 clear options
Research: Users had no reference point between "no idea" and "expert." They didn't know how far to drag.
Decision: Replaced the open-ended drag bar with four discrete options per skill. Users could make a confident choice in seconds instead of guessing.
02
Trust signals added before paywall
Research: All 4 users asked for proof before committing. No social proof = no conversion.
Decision: Added company logos, real testimonials with photos, and a roadmap preview to the signup and pre-payment screens. Users needed to see value before spending $1.
03
Affirmation screens at key moments
Research: Users felt unsure mid-flow and needed reassurance that the platform understood them.
Decision: Added affirmation screens at the start, mid-flow, and right before payment, each personalized to the user's career stage.
04
Testimonials matched to user profile
Research: Generic testimonials didn't land. Users wanted to see people "just like them" who succeeded.
Decision: Built a segmented testimonial system, Breaking In, Growing in Role, Leadership, so the testimonial matched the user's own career stage and struggles.
The Solution
Streamlining Onboarding
I mapped the full onboarding flow and identified where users were dropping off. Added progress indicators so users always knew where they were, simplified multi-step configuration into digestible chunks, and added contextual cues to reduce confusion at key decision points.
Rebuilding Trust on the Landing Page
I redesigned the landing page from the ground up — adding real product screenshots instead of abstract illustrations, testimonials from actual users, partner logos, and concrete numbers showing platform value.
The goal was to answer the user's unspoken question:
"Can I trust this enough to pay?"
Building the Design System
Before scaling to 20+ screens, I built a component library in Figma — buttons, cards, form fields, dashboard modules. This meant every screen felt consistent and handing off to developers was clean and fast.
AI-Assisted Prototyping
Used Bolt to accelerate iteration speed, moving from concept to prototype 4x faster than traditional methods. This was critical for the 48-hour investor demo deadline.
I used Bolt to rapidly generate UI scaffolding from prompts, which meant I could test layout directions in minutes instead of hours. What would have taken 3-4 days of traditional design work got compressed into one overnight sprint.
This wasn't about shortcuts — it was about being a designer who knows how to use the tools available to ship under real pressure.
Understanding the Product Logic
GPT Mapping
I mapped every onboarding screen to what the AI needed as input and what it would give back — so the design was logic-driven, not just visual.
- Target role + resume data → career stage classification
- Years of experience + PM level → roadmap depth and tone
- Struggles + timeline → affirmation content and testimonial matching
- All inputs combined → personalized development plan
Working with the PRD
I worked closely with the PM on the PRD so my designs were technically informed, not just aesthetic.
- Role strata normalization — user's typed goal got classified into one of four career stages: Breaking In, Mid-Career, Senior, or Founder
- Affirmation logic — each stage triggered a different set of affirmation screens, personalized to where the user was in their career
- Testimonial segmentation — testimonials were matched to the user's stage and struggles, not shown generically
- Development plan generation — GPT output structured the roadmap based on all inputs combined
Onboarding Redesign — Three Iterations, One Goal — Getting Users to Trust the Product
Before the Redesign
The original onboarding had the right idea but poor execution.
Skill assessment used an open-ended drag bar with no reference points between "no idea" and "expert." The screens showing current skill level and target outcomes used clashing colors — traffic orange and St. Patrick's green — with no consistency between them. The signup screen had zero credibility signals, and the roadmap preview gave users little reason to believe the $1 payment was worth it.
Overall the UI felt unpolished and not ready for a product asking people to pay.
First Iteration
The first pass focused on foundational UI cleanup and quick UX wins.
Replaced jarring traffic orange and green with subtle, consistent colors across all screens. Swapped the open-ended skill drag bar — which gave users no reference points between "no idea" and "expert" — with four clear options, making self-assessment faster and less frustrating.
Added company logos and testimonials to the signup screen to establish credibility before users committed to anything. Gave the roadmap preview a cleaner structure so users understood the value before spending $1.
Second Iteration
The second pass went deeper on personalization and trust.
Improved testimonial layout with real photos and better UX copy throughout. Added affirmation screens at key moments — start, mid-flow, and just before payment — to maintain momentum without overwhelming the user.
Made testimonials feel more relevant by showing people similar to the user. Consolidated the two "where you are vs where you'll be" screens into one structured, color-coordinated table with clear information hierarchy.
Third Iteration
The final pass refined the trust flow right before the critical payment moment.
Added a dedicated testimonial screen immediately before the payment screen — the highest drop-off point. Removed one affirmation screen that was breaking the flow, keeping only the opening affirmation and letting testimonials carry the weight at the end.
Filled in edge case screens to make the flow production-ready.
User Flows
Two primary flows were mapped to understand the end-to-end experience — the onboarding journey from sign-up to dashboard, and the returning learner path through courses and coaching.
User Flow 1
Three moments in this flow were most likely to cause drop-off, self-assessment, role selection, and pre-payment. Each was redesigned based on interview findings: the drag bar replaced with 4 options, trust signals added at signup, and a roadmap preview shown before the $1 paywall
Returning Learner Flow
The returning learner flow was designed around a key insight, users needed to see where they were and what was next before they'd engage further. The dashboard surfaces progress first, then guides toward courses and coaching, with payment coming only after they've already seen the value.
The Dashboard Redesign
Direction 1 — Courses First
Before committing to a final layout, I explored multiple structural directions to figure out the right information hierarchy for the learner dashboard. The core question was what a user should see first when they log in — and what should live one level deeper.
The first direction led to the main screens — Home, Skill Breakdown, Roadmap/Courses, and Profile — organized around giving learners a clear view of where they are and what's next.
Layouts Explored
Final UI Screens
Direction 2 — Assessment First
The second direction flipped the model. Instead of courses first and assessments second, we explored an assessment-first approach — learning while doing, with courses surfacing contextually on the side rather than leading the experience. This was a directional exploration to test whether that model felt more intuitive for the user.
The product ultimately shifted focus after this point, but these screens represent the core UX thinking behind the learner experience.
The Result
MVP delivered in 48 hours for a high stakes investor demo
20+ screens designed using a scalable and reusable design system
Three onboarding iterations — each driven by a specific gap in trust and usability.
Trust signals redesigned across landing page and onboarding flow
This project taught me that in AI products, the design work isn't just about screens — it's about understanding the logic well enough to design around it.
Learnings
01
Speed and quality aren't opposites
Delivering a full MVP in 48 hours taught me how to ruthlessly prioritize. Not every screen needs to be perfect, just the right screens. Knowing which ones those are, and making that call under pressure, is the actual skill.
02
Trust is a design problem
Users not converting wasn't a marketing problem, it was a UX problem. The landing page wasn't giving people enough reason to believe in the product. Once I treated credibility as a design challenge rather than a messaging one, the solution became clear.
03
AI tools change what's possible
Using Bolt and Cursor wasn't about taking shortcuts — it was about unlocking an iteration speed that a single designer working alone couldn't achieve otherwise. Delivering investor-ready screens in 48 hours wouldn't have been possible without them.
