- Company
- Brainly
- Platform
- Web
- Role
- Senior Product Designer
- Team
- Product, design, UX research, engineering, brand, and QA
- Timeline
- 2 months
- Year
- 2025
About project
Brainly is one of the world’s largest learning platforms with over 300M students globally. The product was evolving beyond its roots in community-driven Q&A toward becoming an AI-powered Learning Companion.
Students came to Brainly for a quick answer, but homework, notes, and exam preparation were spread across different tools. The AI Learning Companion connected those tasks into an ongoing study journey.
As Senior Product Designer, I owned the product design vision, AI chat interactions, and core user journey, and co-defined the beta scope. Working with product, research, engineering, and brand, I took the design from discovery through beta release in two months.
Challenge
Brainly’s goal was to evolve from homework help into a comprehensive AI-powered Learning Companion, trusted for both quick answers and ongoing learning support.
At the product level, we needed to connect Q&A, Homework Sessions, Test Prep, and other tools into a seamless learning journey. Simpler navigation would help students find and use the tools they needed.
The business goal was to improve engagement, retention, and lifetime value by giving students reasons to return beyond a single question. We needed to deliver this new value while limiting negative impact on monetization.
My role & project scope
I owned the product design vision, AI chat interaction design, and the core journey from finding an answer through a Google search to preparing for an exam. I co-defined the beta scope, built prototypes, and partnered on research and validation.
Another designer supported design-system updates for the refreshed branding. I collaborated with the Creative Director and Design System Designer to bring that identity into the product experience.
Two months covered discovery through release. We kept Q&A, introduced Homework Sessions, and integrated Test Prep in the beta. We deferred personalization because of implementation complexity.
Research
Together with the PM and UX researcher, we combined more than ten discovery interviews, AI chat usage analysis, and competitor research. Students used multiple tools for homework, notes, and exam preparation. They wanted more structure and continuity between sessions.
In the month before the project, 30% of students who visited a question page used AI chat. The opportunity was to connect that help with a longer study session.

I conducted competitor research on AI-powered study tools, including ChatGPT, Kami, Quizgecko, AI Tutor, The Drive AI, and Question AI.
Synthesizing the discovery findings with the interface patterns in the benchmarking board pointed to four design opportunities:
- Make the first step clear. Suggested prompts and task-based entry points can help students start with a concrete learning need.
- Give learning a persistent home. Visible history and organized study materials can help students return to their work and continue across sessions.
- Connect answers to a next step. The journey should lead naturally from a question to follow-up help, saved notes, and practice for an upcoming test.
- Balance quick help with structured study. Keep AI assistance easy to reach while making Homework Sessions and Test Prep discoverable within the same experience.

Define
We went through multiple rounds of brainstorming and stakeholder alignment to solve key challenges. We constantly balanced technical constraints, business objectives, new branding and user needs.


Ideation & User testing
I explored different ideas for AI chat interactions, including a side panel and a Perplexity-style chat experience. I initially advocated for the Perplexity-style direction, but evaluating it with the team showed that it would be harder to implement and less scalable for the overall Learning Companion vision.


So I thought about how to bring AI help into Brainly’s existing learning journey instead. The prototype brought AI chat, related Brainly questions, notes, and documents into a Homework Session, with chat presented as one of the tabs. We tested it with five high-school students.
Key findings
The test supported the direction of a connected study session, but exposed a break in how students accessed AI help.
- Students liked having their study materials organized in one place.
- They valued note creation because making notes by hand took time.
- They were confused when chat behavior changed with the selected tab.

Contextual AI that’s always ready to help
After user testing and feedback from project stakeholders, I revised the Learning Companion interaction so chat stayed visible and available for follow-up questions, with context following the student’s place in the workflow. The selected approach was also easier to integrate into Brainly’s existing UX. Students could start with an expert-verified answer and continue exploring through AI suggestions and explanations.

Homework sessions
In a Homework Session, students could work through questions generated from uploaded materials and save study notes. Chat remained available alongside their work, supporting follow-up questions as they moved between tasks.


Test Prep sessions
Based on previously generated worksheets, AI generates practice and builds a plan with daily check-ins, with context-aware support throughout.
- Adaptive revision plans
- Recaps and summaries after sessions
- AI help when students struggle during a session


Interaction design
I used Figma and Vercel v0 to prototype AI chat interactions. Working prototypes let the team review timing and interaction behavior directly. Having the prototypes in code also made handoff easier.


From interaction design to release
I prepared specifications for navigation, AI chat, and question states, then supported engineering with design QA and a change log.
The coded prototypes communicated motion and behavior, while the specifications covered states and layout. Together, they gave engineering concrete references for implementation.


Outcome
We launched the Learning Companion beta for 1% of users over a two-week period, connecting Q&A, Homework Sessions, and Test Prep. The design vision laid the foundation for subsequent releases.
As a result, we saw:
- 20,000
Sessions created
- +8%
Retention
- +10%
AI chat engagement
Release
The beta results gave the team enough confidence to prepare a release for all users. We also began marketing campaigns introducing the Learning Companion through a dedicated landing page, announcement article, and launch video.
Watch the launch video on YouTube
Beyond the beta
After the beta release, I explored a broader vision for the AI Learning Companion: making answers more engaging through interactive widgets.

Tradeoffs
I advocated for simple onboarding or a video explanation supported by help articles, but stakeholders deferred that work for the release. Looking back, I believe existing users faced too many changes at once without enough explanation.
Today, I would make the transition experience part of the release scope: introduce changes gradually, explain each new feature, and check comprehension and adoption before expanding the rollout.
The project also showed me the value of design engineering within a team. Coded prototypes made interaction details tangible, but helping students understand the new workflow needed the same attention as building it.
What I learned
AI is reshaping not just products, but how we build them.
This project introduced a different way of working. We operated as a tight, cross-functional squad: designer, CTO, PM, researcher, and developer, collaborating in real time with minimal overhead. It felt like a true co-creation process, where decisions happened fast and ideas evolved through shared ownership.
Big kudos to the team ❤️
CTO Bill, PM Aga, designers Fabio, Ela, and Grzegorz, UX researcher Weronika, full-stack developer Przemek, and copywriter Dhanesh.