Get in touch
Nord Anglia Education wanted to change how students develop critical thinking — not through another edtech add-on, but by making self-reflection feel like a real conversation. The challenge? Traditional approaches lacked engagement, and teachers had no scalable way to see how students actually think. Together with Nord Anglia, Master of Code Global built Mira — an AI-powered assistant that guides students through reflective dialogue using voice and natural conversation, giving educators real-time insight along the way. Here's how we made it happen.
Mira consistently outperforms traditional learning tool’s baseline across all 7 metacognitive indicators, even with a smaller initial sample size. This demonstrates stronger contextual awareness, self-regulation, and planning skills.
Nord Anglia was exploring how to support students in developing metacognitive skills – the ability to reflect on and understand their own thinking – without adding extra pressure on teachers or disrupting existing learning processes. Traditional approaches relied heavily on manual observation and feedback, making it difficult to scale personalized guidance across classrooms.
At the same time, introducing AI into an educational environment brought an additional layer of complexity. The solution had to be engaging enough for students, intuitive for everyday use, and, most importantly, built with strict safety and privacy considerations. Any interaction needed to be carefully moderated, ensuring that sensitive topics were handled responsibly while still allowing students to express themselves freely.
For teachers, the challenge was equally significant. They needed clear, structured insights into student reflections without spending hours reviewing conversations. The goal was to provide meaningful insights and actionable feedback tools that enhance, rather than overload, their workflow.
Master of Code Global helped Nord Anglia develop Mira – an AI-powered reflection platform combining conversational intelligence, voice interaction, and teacher-focused analysis
To bring Nord Anglia’s vision to life, we designed and delivered Mira – a Proof of Concept solution that transforms reflective learning into an interactive, AI-driven experience. The platform enables students to engage in guided conversations with an intelligent assistant, while equipping teachers with clear, actionable insights into each student’s progress.
At its core, Mira blends multiple AI capabilities into a single, cohesive environment, making reflection more engaging for students and significantly more manageable for educators.
AI-guided reflection conversations: Students interact with an AI assistant in both text and voice formats. The system dynamically adapts its questions based on responses, encouraging deeper thinking and gradually guiding learners toward higher levels of metacognitive awareness.
Voice interaction with an engaging 3D avatar: To create a more immersive experience, the solution includes a 3D avatar that communicates with students via voice. Powered by Azure speech recognition and Heygen animation technology, the platform provides accurate input capture and natural interaction, while the “hold-to-talk” feature maintains clarity even in noisy classroom environments.
Image-based reflection support: Students can upload images in their portfolio as part of their reflection process. The AI analyzes visual content and incorporates it into the conversation as needed, expanding the ways students can express and explore their thoughts.
AI-powered analysis for teachers: The platform automatically processes each interaction to generate concise summaries, highlight key metacognitive indicators, and assess the depth of student reflection. This allows teachers to quickly understand progress without reviewing entire conversations.
AI-assisted feedback generation: Mira drafts personalized feedback based on each session, giving teachers a strong starting point that can be reviewed, refined, and shared with students – saving time while maintaining human control and touch.
Role-based access and data control: The system is structured with clear access boundaries: students only see their own reflections, teachers access data within their classes, and administrators manage users without visibility into personal content — offering privacy by design.
Built-in multilingual experience: The platform supports both English and Spanish, with a fully localized interface. During the pilot phase, we intentionally tested multiple language options to better support multilingual students, enabling them to engage and reflect in their preferred language. The architecture is designed for seamless expansion into additional languages as needed.
Multi-layered safety and moderation system: Mira incorporates advanced content filtering and internal moderation logic to provide safe interactions. The system detects sensitive topics, manages off-topic conversations, and flags potential concerns for teacher review – balancing open dialogue with responsible oversight.
