What I owned
Every pixel, B2C and B2B: the consumer AI chat, enterprise dashboards, payment flows, onboarding, and partner portals for three universities.
Back to work01 / 05
Product Design · AI · EdTech · 2024
A Forbes startup had a design problem. I was the only designer.
Solo designer for a Forbes 30 Under 30 startup, redesigning AI chat, fixing payment drop-offs, and building a multi-brand design system. All at once.
Four products bleeding at once. One designer. No system, no playbook, no safety net.
%20%
engagement lift after landing redesign
preview soon
- Role
- Solo Product Designer, all B2C + B2B
- Duration
- Jul 2024 – Feb 2025
- Platforms
- Web · Mobile · Admin · Partner portals
- Industry
- EdTech · AI
The setup
Tutoruu builds AI tools that personalize learning for university students, and white-label platforms for institutional partners. Not "I contributed to some screens." Every pixel: consumer AI chat, enterprise dashboards, payment flows, onboarding, partner portals for three different universities. One designer. Thousands of users. Real stakes.
Role & constraint
The constraint
No design system, constant pivots, and a market the global playbook ignores, Egyptian students who pay with InstaPay and Fawry, not credit cards.
The problems
What was actually broken.
Problem 01
The landing page was bleeding users
Hotjar heatmaps and Mixpanel data told a clear story: high bounce, low conversion. The page tried to serve B2B and B2C at once, and failed both. Users didn't know if they were a student or an institution.
Problem 02
Ruu AI wasn't delivering its core promise
Users engaged with the AI chat, then left frustrated. Session recordings showed why: they didn't know how to prompt it. Vague questions in, poor results out, churn. The AI was fine. The UX around it wasn't.
Problem 03
The payment page had a massive drop-off
Sales flagged it, Mixpanel confirmed it. Users reached the payment screen and stopped. The platform only accepted credit cards, but the demographic was Egyptian students, low card penetration. The friction was structural, not superficial.
Problem 04
No design system = everything was slow
Multiple products, multiple teams, zero single source of truth. Every brand update meant manual rework across every product.
How it went
From signal to shipped.
Listen
Hotjar heatmaps, Mixpanel funnels, and session recordings named the real problems, before I touched a pixel.
Reframe
Turned every metric drop into a UX hypothesis, not a visual tweak. The AI didn't need fixing; the path to it did.
Ship & iterate
Card sorting → redesign → usability testing → two iteration rounds, all in production with real users.
The solutions
What I actually did about it.
01Solution 01 / 04
Landing page redesign
Heuristic audit → marketing alignment → competitor analysis → card sorting with 10+ users → redesign → usability testing → two iteration rounds.
↑ 20% engagementlonger sessionsclean B2B/B2C split
preview soon
02Solution 02 / 04
Ruu AI: Presets & Mini Ruu's
Users lacked prompting skills, not willingness. The fix wasn't a better AI. It was a UX bridge.
Designed pre-built prompt templates delivered through low-friction modals, so users got value without knowing how to talk to the AI.
More users getting useful answersfewer support tickets
preview soon
03Solution 03 / 04
Payment flow overhaul
Students don't use credit cards. They use InstaPay and Fawry.
Added both rails, then restructured pricing tiers using anchoring and the decoy effect.
Fewer drop-offs at the pay wallmore completed checkouts (sales-confirmed)
preview soon
04Solution 04 / 04
Multi-brand design system
Built a variables-based token architecture with multi-brand theming, exported to JSON for dev handoff, validated in Storybook.
Brand changes in minutes, not daysone source of truth
preview soon
By the numbers
%20%
engagement lift
3
university partner portals
2
new payment rails added
1
designer, every pixel
The AI was fine. The UX around it wasn't.
Gallery
A closer look.
The screens behind the story. High-fidelity shots landing soon.
tutoruu/0116:9
tutoruu/024:3
tutoruu/034:3
tutoruu/041:1
Reflection
What I'd carry forward
The design system should have come first. Once it shipped, brand changes dropped from days to minutes and every product got more consistent overnight. Next 0→1: system before screens.
Next case · 02
Taa Marbuta
265 screens. One mission: a health app that's actually hers.