Designing a scalable, AI-enabled education marketplace

Education services are fragmented. Parents struggle to find and trust tutors, tutors face complex onboarding, and admin teams rely on manual workflows. Existing platforms lacked personalization and couldn't scale.

How might we design one system where parents, tutors and admins each have a different mental model, yet the experience stays consistent and intuitive?

Research identified three groups with distinct needs:

• Parents and learners want trust, clarity and ease, with a low tolerance for friction.
• Tutors want visibility and predictable income, and use the app frequently.
• Admins are power users who need high-frequency actions and zero errors.

“AI features fail when users don't understand why something is suggested.”

The strategy rested on two principles: clarity over completeness and trust-first UX. Progressive disclosure handled complexity, and AI always kept a human in the loop.

Marketplace

Modular tutor cards with a strong hierarchy put expertise and reviews first. Filters reveal themselves progressively so parents aren't overwhelmed, and a guided-match flow asks structured questions upfront to reduce decision fatigue.

Sapience One app

For parents and learners, the goal was minimal cognitive load: calendar-based scheduling people already understand, and clear micro-interactions to confirm critical actions like booking and payment.

Tutor app: simplified onboarding with clear progress indicators, and a dashboard focused on earnings and upcoming sessions.

CRM and admin dashboard: role-based dashboards reduced context switching, and data visualization replaced raw tables for monitoring platform health. Predictive analytics surface personalized insights for learning center managers.

AI assists decisions; it doesn't make them. Every AI surface was designed around explainability and confidence.

• Personalized suggestions come with a reason, such as "Matches your schedule".
• Admin insights flag potential issues but leave the final call to a person.

Recaps give parents and tutors key takeaways, progress tracking and next steps after every session, making learning visible and building trust in the platform.

Engagement rose about 35% across platforms and onboarding activation improved about 40%. A shared design system meant fewer support tickets and faster development iterations, with 98% design-to-code accuracy.