Case Study · 8 min read
73% of students feel “completely overwhelmed” by college admissions, yet most tools treat this $2.9B market like a search problem — not an emotional one. When Reach Best’s CEO showed me screenshots of students crying in our feedback, I knew we had to redesign how AI talks to stressed teenagers. Four months later, our empathy-first platform was supporting 10,000+ students across 40 countries.
ROLE
Product Designer
TIME
8 Months
TEAM
Ryusei
CEO
Hritik
Full Stack Developer
Soufiane
Frontend Developer
"Every app gives me the same list of schools, but none of them know that I cry myself to sleep worrying about my future."
Research
Insights
Archetypes
Prototyping
Testing
Launch
I mapped out the competitive landscape:
Naviance → Trusted but outdated; disengaging for students.
MaiaLearning → Widely used, but overwhelmingly counselor-focused.
Unifrog → Rich database, but complex navigation and poor student engagement.
From my research, I created student personas that reflected not just academic goals but also emotional
Persona 1
The Focused but Overwhelmed Student
Riya Sharma
17F, High School Senior
Goals:
Discover universities that fit her strengths, interests, and ambitions.
Get clear guidance on essays, deadlines, and requirements.
Stay motivated while reducing stress and confusion.
Challenges:
Feels paralyzed by too many options and scattered information.
Struggles with anxiety about rejection and self-doubt.
Easily overwhelmed by excessive data and conflicting rankings.
Limited access to personalized counselor time.
Behaviors:
Researches online late at night, mostly on mobile.
Creates “dream lists” of colleges but struggles to prioritize.
Relies on peer reviews and AI tools for instant feedback.
Motivated by progress tracking and milestone achievements.
Persona 2
The Supportive but Overloaded Counselor
Anil Verma
42M, School Counselor (500+ students)
Goals:
Help students efficiently discover and apply to best-fit universities.
Manage and track student applications in one place.
Provide data-driven, personalized guidance without burnout.
Challenges:
High student-to-counselor ratio leaves little time per student.
Overwhelmed by repetitive queries from hundreds of students.
Lacks easy-to-use tools to organize and visualize applications.
Behaviors:
Prefers desktop tools for managing dashboards and data-heavy tasks.
Uses mobile notifications only for quick updates.
Relies on dashboards that give quick snapshots of student progress.
Values AI tools that reduce repetitive tasks and allow deeper mentoring.
Persona 3
The Focused but Overwhelmed Student
Lina Chen
17F, High School Senior
Goals:
Secure admission to a reputable US university for Computer Science.
Find reliable scholarship opportunities to make studying abroad affordable.
Gain clarity on SAT, TOEFL, and application essay expectations.
Challenges:
Confused by inconsistent information across forums, agencies, and websites.
Feels isolated since few peers in her school are applying abroad.
Anxious about cultural differences and fitting into a new environment.
Struggles to compare universities beyond rankings and brand names.
Behaviors:
Joins online communities of international applicants for advice.
Watches YouTube videos about campus life and admissions tips.
Keeps multiple spreadsheets to track deadlines, tests, and documents.
These personas became the foundation for testing how empathy-driven AI could respond differently to different emotional patterns.
I identified recurring emotional patterns and designed AI mentors as archetypes inspired by real student journeys.
In testing with 12 students, 9 preferred personality-driven AI over generic responses. This validated that emotion could be a differentiator.
My hardest design decision: Creating AI personalities that felt authentic without being manipulative.
What I owned:
All user research
Persona development
UX/UI design
Prototyping
Full design system (built from scratch, no UI libraries)
My early prototypes came across as “chatbot therapy.” One student said:
"This feels like talking to a guidance counselor robot."
I realized I was building AI that analyzed problems instead of AI that shared struggles. Pivoting toward storytelling and emotional resonance improved engagement dramatically.
AI Mentors
I created scalable story-driven personalities dozens of them — like Maya the transfer Student, Alex the STEM Explorer, and Jordan the International Applicant. These AI mentor personalities validated feelings first, then offered practical guidance.
AI Mentors - Chat Interface
Essay Review
Feedback system that helped students refine their own essays instead of outsourcing them. (Research showed 84% wanted to “own their stories”).
AI Essay
RB Search
Goal-based recommendations that curated schools based on fit, not rankings. AI explained why each school matched.
RB Search
Smart Notes
A Medium-style editor for capturing insights, deadlines, and reflections—keeping admissions knowledge in one place.
Smart Notes
Campus Exploration
Integrated Google Earth with game-like controls, turning research into an interactive, tactile experience.
Campus View
Progress Dashboard
Real-time visibility into each student’s journey, including:
Which universities they applied to
Whether essays were submitted
If recommendation letters were uploaded
This gave counselors a clear picture of student progress without chasing updates over email.
Counselor Students View
Counselor Students View
Aggregate Insights
Trends across the entire student body, such as how many applied to a specific university or country. This helped counselors benchmark outcomes, allocate time, and prepare resources for the most popular destinations.
Counselor Analytics
Smart Filters
A quick way to surface disengaged or at-risk students, so counselors could focus their limited time where it mattered most.
Counselor Analytics
Multiple schools became paying customers within 6 months
Ethical positioning improved conversion rates significantly
Organic adoption reduced reliance on marketing spend
10,000+ students across 40+ countries
Students averaged 18 messages per mentor conversation (internal analytics)
In post-conversation surveys (n=643), 72% reported feeling less anxious
34% returned within 48 hours (Hotjar analytics)
5M+ reach across TikTok and Instagram
Invited to present at the Council of International Schools (only 2nd company after Duolingo)
Positive feedback from educators highlighted our ethical design stance
Featured in a Mexican news channel and the San Francisco Chronicle for our empathetic AI approach
From Solo Designer to Systems Thinker and Story Architect
WHAT I LEARNED:
⚡
Even the smartest tools fall flat if they don’t feel human. Empathy drove more impact than algorithms ever could.
⚡
With no UI kits or libraries, I was forced to think in systems and build a lean, scalable design language from the ground up.
⚡
By choosing not to let AI write essays, we ended up building tools that supported deep thinking—and earned lasting trust.
⚡
Ethical design choices, when aligned with user values, can become your strongest competitive advantage.
