My PhD research on student-centred eLearning revealed a fundamental challenge: how do we provide sophisticated support for learning without removing the struggle that builds understanding? Today, as AI transforms education, this question has become urgent. The answer lies not in AI that provides answers, but in AI agents that scaffold thinking while preserving student agency.
Student-centred learning has always been about putting learner needs, interests, and agency at the heart of education. But implementing this philosophy at scale has been limited by the practical constraints of traditional classroom management. AI agents offer a path forward—technology that can provide individualized support without sacrificing the core principles that make student-centred learning effective.
What My Research Revealed About Student Agency
Through extensive observation of student-centred eLearning environments, my research consistently showed that the most effective learning happened when students retained control over their learning process. This wasn't just about choice in topics or pacing—it was about preserving the cognitive work that builds understanding.
Students learned best when technology provided scaffolding rather than shortcuts. The difference is crucial: scaffolding supports the learning process while maintaining challenge, while shortcuts bypass the process entirely.
Key Findings from PhD Research
Student-centred learning environments work best when they:
- • Preserve productive struggle: Students need to grapple with concepts, not receive pre-digested answers
- • Maintain learner agency: Students should control the pace and approach to their learning
- • Provide just-in-time support: Help should arrive when needed, not before it's requested
- • Encourage metacognition: Students need to understand how they learn, not just what they learn
- • Support multiple pathways: Different students need different routes to the same learning goals
"The goal of student-centred learning isn't to make learning easier—it's to make learners more capable. This requires preserving the cognitive work while removing unnecessary barriers." - From my PhD dissertation
The Problem with Current AI in Education
Most current AI educational tools violate the core principles of student-centred learning. They optimize for efficiency over agency, answers over understanding. While these tools can help students complete assignments faster, they often undermine the very processes that build lasting learning.
How Current AI Undermines Student-Centred Learning
- • Replaces thinking with outputs: AI writes essays, solves problems, and provides answers without student cognitive engagement
- • Removes productive struggle: Students get immediate answers instead of working through challenges
- • Standardizes approaches: One-size-fits-all solutions ignore individual learning differences
- • Diminishes metacognition: Students focus on using AI tools rather than understanding their own learning
- • Creates dependency: Students become reliant on AI for tasks they should learn to do independently
I've observed this pattern repeatedly: students using AI tools become more efficient at completing assignments but less skilled at independent thinking. They can produce higher-quality outputs in the short term but struggle with novel problems that don't fit their AI-assisted patterns.
AI Agents: A Different Approach
AI agents represent a fundamentally different approach to educational technology. Rather than providing answers or completing tasks for students, agents act as sophisticated learning partners that understand context, support process, and preserve student agency.
The Agent Difference in Practice
Traditional AI Tools
Student Request: "Help me write an essay about climate change"
AI Response: [Produces complete essay]
Result: Student gets essay but learns nothing about research, argumentation, or writing
AI Agent Approach
Student Request: "Help me write an essay about climate change"
Agent Response: "What specific aspect interests you most? Let's start by helping you develop a focused thesis..."
Result: Student develops research, critical thinking, and writing skills through guided process
Agent Characteristics That Support Student-Centred Learning
Process-Oriented Support
Agents focus on improving learning processes rather than providing learning products. They help students develop better research strategies, clearer thinking patterns, and stronger problem-solving approaches.
Adaptive Scaffolding
Agents adjust their support based on student needs and progress. They provide more structure for struggling learners and more independence for advanced students, always working to gradually reduce support as competence grows.
Metacognitive Enhancement
Agents help students understand their own learning processes. They ask reflective questions, highlight thinking patterns, and help students develop self-regulation skills that transfer to new contexts.
Context Awareness
Agents understand individual student contexts—learning preferences, prior knowledge, current challenges, and long-term goals. This allows for truly personalized support that honors each student's unique learning journey.
How Zaza Applies These Principles in Teacher Tools
While our current tools focus on supporting teachers rather than students directly, they're designed with the same student-centred principles that guided my research. We believe that better teacher tools lead to more student-centred classrooms.
Zaza Teach: Preserving Teacher Agency
Just as students need agency in their learning, teachers need agency in their teaching. Zaza Teach provides lesson planning support that preserves teacher creativity and professional judgment while reducing administrative burden. The AI scaffolds the planning process without dictating teaching approaches.
AutoPlanner: Differentiation Support
AutoPlanner helps teachers create differentiated learning experiences that honor student-centred principles. It suggests multiple pathways to learning goals, varied assessment strategies, and adaptive pacing—all aligned with curriculum standards while preserving student choice and agency.
The Future: Direct Student-Agent Partnerships
Looking ahead, I envision AI agents working directly with students as learning partners. These agents will understand individual learning profiles, adapt to different thinking styles, and provide just-in-time support that maintains challenge while reducing unnecessary friction.
Vision for Student-Agent Learning Partnerships
Personalized Learning Coaches
Agents that understand each student's learning style, pace, and interests, providing individualized support that adapts in real-time.
Thinking Process Mentors
Agents that help students develop stronger problem-solving strategies, critical thinking skills, and metacognitive awareness.
Collaborative Learning Facilitators
Agents that support group work, peer learning, and collaborative problem-solving while maintaining individual accountability.
Adaptive Assessment Partners
Agents that provide ongoing, formative assessment that informs learning rather than just measuring it.
The Implementation Challenge
Building agents that truly support student-centred learning requires more than sophisticated AI—it requires deep understanding of learning theory, cognitive development, and the complex dynamics of classroom environments.
Critical Design Considerations
- • Maintaining Challenge: Agents must provide support without removing productive struggle
- • Preserving Agency: Students must retain control over their learning decisions
- • Developing Independence: Support should gradually decrease as competence increases
- • Honoring Diversity: Agents must work with different learning styles, backgrounds, and abilities
- • Supporting Teachers: Agents should enhance rather than replace teacher expertise
This is why my PhD research remains so relevant to our work at Zaza. The principles we discovered about effective student-centred learning provide the foundation for building AI agents that truly serve learning rather than replacing it.
From Research to Revolution
The transition from traditional educational technology to agent-native learning support represents more than technological advancement—it's a return to fundamental principles of effective education, enhanced by sophisticated AI capabilities.
My PhD research showed that student-centred learning works when it preserves agency, maintains challenge, and supports individual pathways to understanding. AI agents offer the first realistic path to implementing these principles at scale while maintaining the human relationships that make learning meaningful.
The Promise of Student-Centred AI
The next era of educational technology will be defined by:
- AI that enhances rather than replaces student thinking
- Technology that adapts to learners rather than forcing learners to adapt to technology
- Tools that preserve the challenge and agency essential for deep learning
- Agents that understand context, support process, and honor individual differences
- Educational AI that makes student-centred learning practically achievable at scale
The question isn't whether AI will transform education—it already is. The question is whether we'll use AI to create more student-centred learning environments or more efficient content delivery systems. The choice we make will shape how an entire generation learns to think, question, and grow.
My PhD research convinced me that the most effective learning happens when students maintain agency over their own thinking. AI agents offer us the tools to honor this principle while providing the sophisticated support that makes challenging learning accessible to all students. The era of agents, not answers, is just beginning.