AI-Powered Homework Help
I spearheaded research for Google Lens’ AI-powered Homework Help feature, guiding product strategy, design, and model tuning—from early co-design with students to real-world testing across markets—ensuring the experience was delightful, trustworthy, and truly helpful.
Featured press: Back to school season easier with these AI features, Lens EDU in Indonesia
The goal
Students (especially GenZ) don’t want to hunt for homework help in a sea of web results.
How can we leverage visual intelligence and generative AI to reduce friction, save time, and help students learn better by providing step-by-step, contextually rich support over static results?
The approach
Early Concept & Co-Design:
Hands-on co-design with high school students to define what problems mattered most, ideate solutions and prioritize features—from solving a simple problem to studying a complex topic.
Facilitated educator focus groups to address ethical concerns, foster support, and shape a solution that balanced student autonomy with pedagogical responsibility.
Usability & Comprehension Testing:
Led iterative concept and usability testing to ensure users could successfully complete their homework help journey—from image capture to comprehension and trust of AI-generated explanations.
Identified and resolved UX friction points tied to scanning fidelity, user control over outputs, and accessibility for students with different learning needs.
Launch Readiness Assessment:
Ran a nationwide diary study to explore real-world use, track trust and perceived accuracy in AI outputs, and assess how well the tool supported meaningful learning (vs. simply copying answers).
Mapped key sentiment dimensions—accuracy, helpfulness, and differentiation—surfacing actionable insights on hallucination tolerance, misconceptions, and possible barriers to sustained engagement.
Post-launch: Scaled Market Insights
Deployed large scale multi-market surveys to evaluate satisfaction, discover top pain points, and identify growth opportunities across key regions.
Segmented findings to guide regional feature prioritization and quality tuning for top education markets—laying groundwork for sustainable adoption at scale.
The Impact
Contributed end-to-end research support for the 0→1 launch of Google Lens’ AI Homework Help, serving millions of students globally with visual-first, AI-supported academic assistance
Grew education as a core vertical within Lens, evolving it from simple OCR-based search to one of Google’s first generative AI education experiences—setting a foundation for future-visioning with multimodal experiences.
Established cross-functional research principles and connections that continue to inform AI ethics, comprehension guardrails, and model behavior tuning for learning experiences across Google.