The Client: A Digital Language-Learning Company

The client operates in EdTech and provides digital language-learning experiences. It saw an opportunity to strengthen retention by shifting the product focus from module completion toward habit formation and learning growth through Adaptive Learning.

‍

The Mission: Launch Adaptive Learning to Support Retention

Background

Over six months, Product People took the Adaptive Learning MVP from fragmented requirements to launch, then defined the next phase of personalized review, progression, and practice capabilities.

Why Product People Were Needed

Delivering a coherent adaptive experience required alignment across Learning Science, Data Science, Data Engineering, Engineering, and Product Design.

• Retention strategy centered on habit formation and learning growth
• Personalization that balanced pedagogy with algorithmic decisions
• Release readiness across experience, data, and system performance

‍

Our Main Quest: Strengthen Retention Through Adaptive Learning

Turning Retention Strategy into Adaptive Learning

Problem: The client saw an opportunity to strengthen retention by shifting from module completion toward habit formation and learning growth. The challenge was to translate this goal into a coherent, personalized learning experience without compromising pedagogical quality or system performance.

What we did:

• Launched adaptive review by translating learner mastery and difficulty-scoring data into a personalized, learner-facing experience.
• Defined the personalization algorithm and UX with Data Science, Data Engineering, Learning Science, Engineering, and Product Design.
• Built a learner-need-first roadmap from onboarding through retention, informed by competitive research across established language-learning products and emerging AI tutors.
• Moved the MVP through end-to-end testing and reduced latency to 5 seconds, meeting the required performance benchmarks.
• Protected learner progress through a learning-path change strategy that preserved completion after content edits.
• Scoped level transitions and practice modules for Phase 2.

Outcome: Launched the Adaptive Learning MVP with a personalized review experience, 5-second latency, and a defined Phase 2 roadmap.

Scaling AI Content Without Compromising Quality

Problem: The client needed to scale AI-generated learning content without compromising pedagogical quality. Learner reports and exercise rejections contained valuable signals, but there was no repeatable path to convert them into regeneration improvements, reviewer decisions, and actionable Engineering work.

What we did:

• Turned more than 600 exercise-rejection reasons into actionable regeneration inputs for V2 content generation.
• Defined content-review specifications and prototypes with Engineering to streamline content QA.
• Created a repeatable path from learner flags to resolution through triage and reviewer-flow prototypes.
• Scoped requirements and prototypes for bulk assignment to make reviewer operations more scalable.

Outcome: Established a repeatable quality loop connecting rejection insights, reviewer workflows, regeneration improvements, and actionable Engineering work.

‍

🏆 Mission Achievements: Delivered Outcomes

✅ Reframed retention – Shifted the core problem from completion volume to habit formation.

✅ Launched personalized review – Translated mastery and difficulty-scoring data into a learner-facing adaptive review experience.

✅ Reached release performance – Optimization reached 90%, and latency fell to 5 seconds.

✅ Scaled the quality loop – Connected exercise-rejection insights, reviewer workflows, and Engineering actions into a repeatable content QA process.