We are building the adaptive learning platform for all learning. We are building the adaptive learning platform for passing exams.

Today, learners have access to more educational content than ever before. The internet has made knowledge widely accessible, online courses make structured teaching available at scale, and AI can increasingly answer questions, generate exercises and explain almost any subject on demand.

But more content creates a new problem. Deciding what to do next. At any point, a learner could read another explanation, watch another video, practise a new topic or revisit something they learned before. The right choice depends on what they already understand, where their knowledge gaps are and what they are trying to achieve. Without that context it’s hard to focus and progress.

Great teachers and tutors solve this problem by building an understanding of the learner over time. They act as a guide, helping identify gaps and misconceptions, intervening where necessary and adapting the path as the learner progresses. Unfortunately, this level of support is often inaccessible as tutoring is expensive and teachers are stretched.

We believe there is an opportunity to make that kind of continuous, individual guidance available at scale. The missing piece is a persistent understanding of the learner: one that develops across interactions and over time, understands both the individual and what they are trying to learn, updates itself from evidence, and uses that context to continually guide the learner towards their goals.

This is our plan to build an adaptive platform for learning:

1. 1Create a generatable subject knowledge graph.
2. 2Develop an advanced learner memory.
3. 3Build an incredible recommendation engine.

The recommendation engine connects what the platform understands about the subject with what it understands about the learner to decide the most useful next action. It might introduce a new concept, diagnose a missing prerequisite, change an explanation, provide practice, revisit knowledge before it is forgotten or move on once there is enough evidence of mastery.

This creates a learning loop that can improve over time. Every intervention gives us evidence not only about the learner, but about what works, for whom and under which conditions. As more learning loops are completed, the platform should become better at selecting the interventions that lead to lasting progress.

Our first product will be focused on secondary education for students aged 11-18. Secondary education provides a rigorous environment in which to develop the platform. Subjects have structured curricula and measurable outcomes, but every learner arrives with different knowledge, misconceptions and goals. Supporting them effectively requires the platform to understand both what must be learned and the individual path each learner should take.

While our vision is much larger, building an excellent secondary education product requires us to solve the core problems at the heart of the platform. If we can prove those foundations we will have the pieces in place to support the second and third parts of the plan. We'll have more to say about those later.