Back to work
AI Product Mobile · iOS UX Architecture
Mila AI — Your Personal AI Language Teacher

Redesigning a web-based platform into an AI-driven mobile app — reducing friction and making personalisation visible to the user.

Role
Product Designer / UX Architect
Scope
End-to-end product design
Platform
iOS mobile app
Year
2024
Mila AI — app screens overview
85%
Usability improvement in structured testing
1 tap
To resume a lesson — reduced from 4+ navigation steps
Visible AI
Personalisation surfaced to the user — not hidden in the backend
Overview

Mila AI had a real problem: the AI was working, but nobody could see it. Users navigated through multiple menus to reach a single lesson, lost their place constantly, and had no sense the product was adapting to them. The tech was there. The experience wasn't.

I led the end-to-end UX and UI for the iOS app — rethinking how progression, personalisation and continuity should work for someone learning on their phone in short, on-the-go sessions.

The challenge
"A language learning app only works if people come back tomorrow. Mila had the content and the AI — but the mobile experience was an afterthought built on top of a desktop product. Users dropped off before they ever saw the value."
Problems identified
  • Multiple menus to reach a single lesson
  • No learning roadmap or visible progression
  • Users lost context after every interruption
  • AI personalisation completely invisible to users
  • UI designed for desktop, not mobile sessions
Design goals
  • One tap to resume where you left off
  • Clear learning roadmap with visible progress
  • Personalised onboarding from first launch
  • AI adaptation surfaced as a visible feature
  • Experience built for short, interrupted sessions
Mila AI — personalised onboarding flow
Onboarding flow

Users define goals, level and target language on first launch. The system adapts immediately — AI goes from a hidden backend process to a visible feature users can see and trust from day one.

Design strategy
Strategy 01
Mobile-first architecture

Rebuilt the IA around quick access and clear progression. Fewer steps to the lesson, a clearer sense of where you are in your journey.

Strategy 02
Personalised onboarding

Users define goals and level on first launch. The system adapts immediately — creating ownership and demonstrating AI value from the very first session.

Strategy 03
Adaptive learning roadmap

A dynamic roadmap unlocks based on progress — turning AI adaptation from a backend process into something the user can see, feel and trust.

Strategy 04
Resume & continuity

Progress saves automatically. "Continue Learning" is the primary action on every return. Users re-enter without losing context — one tap to resume.

Mila AI — adaptive learning roadmap
Learning roadmap

A dynamic roadmap unlocks content based on performance. Users can finally see where they are, what they've completed, and what's next — making progression tangible and motivating return.

Outcome
✅
85% usability improvement in structured testing with early users — measured directly against the previous web experience.
✅
Progression clarity resolved — users knew exactly where they were, what they'd done, and what to do next without any guidance or prompts.
✅
AI perceived as visible and personal — users described the app as feeling like it was "learning them", rather than serving generic content.

Building an AI-powered product?

Making AI feel useful and visible to non-technical users is one of the hardest UX challenges. It's one I specialise in.

Let's talk