Mualim — Hafsah Zulafiqar

The brief

Mualim needed an AI tutor for kids that felt natural, not bolted-on, for the Saudi Arabian market. I was brought in as product design consultant, owning the work end to end: research, strategy, prototyping, and shipping.

Research first

Designing for kids means designing for two audiences at once — the child actually using the product, and the parent or guardian deciding whether to trust it with their kid's learning. My UX research focused on understanding how each side needed to experience the product differently: what makes an AI tutor feel safe, engaging, and genuinely useful to a child, and what makes it feel trustworthy and transparent to the adult overseeing it. That split informed every decision that followed.

Designing for a child's attention and a parent's trust, at the same time — in a single interface.

The dashboard for parents.

One dashboard to check on their kids progress.

Design exploration — and what didn't work

We explored different design languages and approaches to see which would resonate with kids aged 5 to 15. We needed something that reinforced learning with positive emotions, without overcomplicating the experience.

Mualim exploration — early challenges list layout

Mualim exploration — early challenge task screen

Mualim exploration — challenge task variant

Mualim exploration — early tasks list layout

Attempting challenges

Challenges were designed to be easy to navigate, reduce the cognitive load on students, and encourage them to try again if they got stuck — with the option to easily ask the AI tutor for help along the way.

AI learning

This was designed around an unconventional approach to education — kids didn't need to memorize concepts, but to learn, unlearn, and relearn through iteration and goal-based learning.

Strategy

With the research in hand, I shaped the product strategy around how an AI tutor should actually behave with a young learner: how it introduces itself, how it paces a lesson, how it responds when a child gets something wrong, and how it surfaces progress back to a parent without making the experience feel like it's being graded. My goal was a strategy that treated "AI-driven learning" as a promise to make good on through interaction design, not just a label on the product.