Great AI tutoring starts with understanding how people learn.
In 2026, AI pioneer Andrew Ng introduced LearnVector, an AI-native education company with a vision of transforming traditional one-to-many education into personalised, one-to-one learning.
His vision reflects an exciting shift in education: AI is making individualised learning support more accessible than ever before.
But this also raises an important question:
If AI makes personalised tutoring possible at scale, what makes that tutoring genuinely effective for learning?
AI can explain concepts, generate exercises, simulate conversations, and provide instant feedback. Yet more information does not automatically mean better learning. Nor does personalisation necessarily guarantee that learners develop deeper understanding or greater independence.
At YNova, we believe the answer begins with learning science.
Decades of research have helped us understand how people build knowledge, develop skills, benefit from instructional support, and gradually become independent learners.
We draw on these insights to guide how we design AI tutoring experiences—combining the possibilities of modern technology with established principles of effective learning.
Here are six principles that shape how we turn learning science into practice at YNova.
1. Make Learning Manageable
Break complex tasks into clear steps, gradually building towards independence.
Learning something new can be overwhelming when we are expected to do everything at once.
Research on scaffolding shows the importance of providing appropriate support and gradually reducing it as learners become more capable (van de Pol et al., 2010).
At YNova, we aim to make challenging learning tasks approachable without removing the challenge of learning itself.
In practice: Our IELTS Speaking tutor guides learners from developing ideas to building vocabulary, constructing sentences, producing connected speech, and finally speaking independently.
This journey is also informed by Levelt's (2000) research on how people produce spoken language.
Rather than asking learners to deliver a complete answer immediately, we help them build the skills to do so.
2. Make Learning Personal
Shape practice around learners' ideas, goals, experiences, and demonstrated needs.
People learn with different knowledge, interests, and experiences.
We believe effective personalisation means recognising these differences, not simply giving everyone a different set of exercises.
Research on culturally relevant pedagogy highlights the value of building on learners' existing knowledge and experiences (Ladson-Billings, 1995).
In practice: When practising an IELTS speaking question, learners first explore what they want to say.
Two learners answering the same question might choose completely different stories. YNova helps each develop the vocabulary and expressions relevant to their ideas.
The aim is to help learners communicate something meaningful to them, rather than memorise a generic answer.
3. Use Bilingual Support Purposefully
Help learners build on what they already know and connect their thinking to a new language.
Learning in another language does not mean starting from zero.
Learners already possess knowledge, experiences, and reasoning abilities developed through their existing languages. Research on bilingualism suggests that these resources can support learning across languages (Cummins, 2000).
In practice: YNova offers a bilingual Teacher Mode, where learners can explore ideas and understand explanations in a familiar language.
When ready, they can practise through English-only Immersive Mode.
Our goal is not to replace English practice with translation. It is to help learners use their existing knowledge as a bridge towards independent English communication.
4. Turn Knowledge into Communication
Create meaningful opportunities to apply, revisit, and transfer what learners know.
Understanding something is not the same as being able to use it.
A learner might know a grammar rule but struggle to apply it during conversation. Another might recognise useful vocabulary but be unable to retrieve it when speaking.
Learning becomes more meaningful when knowledge can be used in real tasks.
In practice: YNova connects guided language practice with independent speaking activities and IELTS Mock Tests.
Our Writing tutor, currently in development, follows a similar philosophy through planning, drafting, revising, editing, and reflection.
Across both experiences, we aim to help learners move from understanding language to using it purposefully.
5. Make Feedback Actionable
Show learners what they do well, what needs attention, and what to practise next.
A score can tell learners how they performed, but not necessarily how to improve.
Research on feedback and language learning highlights the importance of helping learners notice difficulties and understand how to address them (Schmidt, 1990; Lyster & Ranta, 1997).
In practice: YNova's Speaking Mock Test provides AI-estimated scores and feedback across the four IELTS Speaking criteria.
Learners can explore their strengths, identify specific problems, review examples, and see suggestions for improvement.
They can also refer to their speaking transcripts and access bilingual explanations to better understand the feedback.
We want feedback to become the beginning of the next learning opportunity, not simply the end of a test.
6. Connect Learners, Teachers, and Schools
Make independent learning visible and help educators provide more informed support.
Learning happens both inside and outside the classroom.
But teachers cannot always see what students practise independently, where they struggle, or how their performance changes.
Research on formative assessment highlights the importance of using learning evidence to guide teaching decisions (Black & Wiliam, 1998).
In practice: YNova connects student practice with teacher dashboards and learning progress reports.
Teachers can review students' participation and AI-estimated speaking performance, identify areas needing support, and plan more targeted activities.
By reducing repetitive review work, we aim to give educators more time for meaningful student support and teaching innovation.
AI should support teachers, not replace them.
Learning Science, Brought to Life
These six principles guide how we build YNova.
We make learning manageable through structured support, personal through learners' own experiences, and accessible through purposeful bilingual assistance.
We connect knowledge with meaningful practice, turn feedback into opportunities for improvement, and help teachers and schools make better use of learning evidence.
We are starting with English language learning and IELTS preparation, but our vision extends beyond a single examination or subject.
We believe the future of AI tutoring isn't simply about providing faster answers.
It's about helping people develop the knowledge, skills, and confidence to find and communicate their own answers.
That's the kind of learning experience we're building at YNova.
YNova is a multilingual AI tutoring platform providing personalised learning experiences for students and actionable learning insights for teachers and schools.
Black, P., & Wiliam, D. (1998). Assessment and classroom learning. Assessment in Education: Principles, Policy & Practice, 5(1), 7–74. https://doi.org/10.1080/0969595980050102
Cummins, J. (2000). Language, power and pedagogy: Bilingual children in the crossfire. Multilingual Matters.
Ladson-Billings, G. (1995). But that's just good teaching! The case for culturally relevant pedagogy. Theory Into Practice, 34(3), 159–165. https://doi.org/10.1080/00405849509543675
Levelt, W. J. M. (2000). Producing spoken language: A blueprint of the speaker. In C. M. Brown & P. Hagoort (Eds.), The neurocognition of language (pp. 82–122). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198507932.003.0004
Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake: Negotiation of form in communicative classrooms. Studies in Second Language Acquisition, 19(1), 37–66. https://doi.org/10.1017/S0272263197001034
van de Pol, J., Volman, M., & Beishuizen, J. (2010). Scaffolding in teacher–student interaction: A decade of research. Educational Psychology Review, 22, 271–296. https://doi.org/10.1007/s10648-010-9127-6
Share
Try NOVA
NOVA coaches IELTS Speaking in your own language, then steps back as you get stronger.
