An AI-powered platform that helps teachers plan courses, lessons, and assessments. Structured, standards-aligned, and built around how teachers actually work.
Product Designer
EdTech
Teachers spend a surprising share of their time on admin and planning work instead of actual teaching: structuring courses, writing lesson plans, aligning content with educational standards, and creating assessments.
Existing AI tools didn't really solve this. Teachers had to figure out the right prompts themselves, lost track of scattered chat threads, and got unstructured output that still needed a lot of manual reorganizing to fit real teaching standards and practices.
The goal: design a product that lets teachers work with AI in an organized way, grounded in real educational standards and practices, without needing to know how to "talk" to AI.
I ran a competitive analysis alongside deep standards research, then surveyed teachers directly to check both against real classroom needs.
I also surveyed teachers directly to understand their biggest pain points in course and lesson planning: how much time it took, what tools they used, and what frustrated them most about existing solutions. These insights fed directly into feature prioritization and the courses → objectives → units → lessons content structure.
Research pointed to two core needs: a guided structure instead of freeform chat, and content that's standards-aligned by default instead of generic AI output.
Let teachers create high-quality, standards-aligned course content quickly, without needing to know how to prompt an AI?
The answer took shape as a structured, guided workflow: teachers move through a clear content hierarchy, courses, objectives, units, lessons, with AI support built into each step, rather than a blank chat box.
Mapped the full information architecture before any screens: how courses break into units, topics, assessment types, and objectives. This gave the AI-generation features a solid structure to plug into.
The home dashboard organizes a teacher's Courses, Presentations, and Reviews, with a guided "How It Works" section introducing each capability. The layout adapts from a grid on desktop to a stacked list on mobile, keeping the same hierarchy across devices.
Courses are shown as color-coded cards with grade, subject, and institution, letting teachers scan and jump into the right course fast, without digging through folders.
A split-panel interface lets teachers navigate a course's units on one side while editing goals, objectives, and outcomes on the other. It's the same structure from the research phase, now directly editable in the UI.
The same split-panel pattern repeats one level down: a list of lessons on the left, a full lesson plan (introduction, instructions, practice, summary, reflection) on the right.
Quizzes generate straight from a lesson's own objectives and content, with a simple control over question count. Teachers get a ready-to-print assessment aligned to what they already planned, instead of writing questions from scratch.
Instead of a freeform chat, AI edits appear inline and reviewable: teachers see a proposed change, can refine it further, undo it, or apply it directly, with pagination when multiple suggestions are available. This came straight out of research: teachers didn't want to chat with AI, they wanted control over specific edits within their own content.
The interface was built on Material Design 3, customized to the product's needs, including a full state matrix across button and input variants, to keep things accessible and consistent as the product scaled across many screens.
EdSyl launched successfully and was presented at several industry conferences, leading to a collaboration with a university. Teachers reported that the AI-assisted workflow felt comfortable and intuitive to use.
I've been a teacher myself for over 10 years, so this project meant more to me than just another product to design. Watching teachers get back hours they used to lose to paperwork, and knowing that time could go toward actually resting, made every research call and design revision feel worth it.
Access to the survey data got cut short during the project. Next time I'd set up a lightweight, ongoing feedback loop earlier, even informal channels, so research insights stay available through the whole product lifecycle, not just the early discovery phase.