Beyond one-size-fits-all AI — aime™ introduces lesson-specific learning environments

aime today set out the next step in its product direction: lesson-specific learning environments. The premise is simple and, for an industry that has settled on a single chat box for every subject, uncomfortable — one prompt and one slide template cannot teach every discipline well. A great Mathematics lesson does not look, move or interact like a great Geography lesson, and pretending otherwise is the reason most AI in education feels the same after the third example.

The generic AI tutor has a ceiling

Most educational AI products share the same architecture: a user prompt, a generic system prompt, a frontier LLM and a generic slide template. It is a clean design and it produces content quickly. It also produces lessons that all teach in the same voice, present in the same layout, lose subject-specific pedagogy in transit and offer none of the interactive tools that make a subject teachable in the first place.

The gap is visible the moment the subject changes. Mathematics needs step-by-step derivations, graph plotting and geometry tools. Geography needs maps, globes and timelines. Biology needs labelled diagrams and anatomy illustrations. Physics needs simulations, free-body diagrams and animations. Programming needs an editor, a terminal and a way to actually run the code. A single presentation environment cannot deliver any of these well, and stretching it to try is what produces the flat, interchangeable feel of most AI lessons today.

aime's position — many environments, one operating system

aime does not build one AI teacher. It builds many specialised teaching environments on top of one operating system. Every lesson is produced through the collaboration of three components already at the centre of the aime stack: EduRule™, ThinkBook™ and the AIME Renderer. Together they let each subject be taught in something much closer to the way an expert educator would teach it.

EduRule — expert-crafted teaching intelligence, per subject

EduRule is aime's pedagogy-aware decision core. Rather than a single prompt that tries to be everything to every subject, EduRule is a collection of subject-specific rulebooks that define how a discipline should be taught: methodology, learning objectives, explanation flow, concept progression, visual preferences, assessment strategy, questioning style and engagement pattern. Crucially, there is not one EduRule — there are many, each written with subject-matter experts and refined against educational practice. The specific contents and structure of each EduRule are aime's proprietary IP and are shared with partners under NDA.

ThinkBook — a pedagogical knowledge graph, not a content dump

Good teaching is not the retrieval of facts in a pleasant order. It is the navigation of prerequisites, dependencies, hierarchies and the misconceptions students reliably bring with them. ThinkBook is aime's pedagogical knowledge graph — a structured representation of those educational relationships that the agent stack can traverse while building a lesson. Instead of generating isolated explanations, aime walks an educational graph that mirrors how a human teacher sequences instruction. The graph schema, the relationships it captures and the way it is consulted at lesson time are part of aime's proprietary architecture.

AIME Renderer — more than slides

Traditional decks are static. A modern teaching surface should not be. The AIME Renderer is aime's custom presentation engine — a slide runtime, an integrated whiteboard and an interactive lesson environment in one. The renderer lets the AI draw while it explains, annotate diagrams in motion, highlight as it speaks, build equations step-by-step and move between authored slides and freehand teaching without breaking flow. The result is closer to a live classroom than to a presentation file.

Subject-specific presentation environments

Content is not enough on its own. Each subject in aime also receives its own presentation environment — handcrafted along two dimensions that matter for learning.

Visual theme. Each subject has a distinct visual identity, so students recognise the context the moment a lesson opens. Mathematics leans on minimal grids and geometric styling. Biology uses organic colours and scientific illustration. Geography is built around earth tones and map-centric layouts. Programming takes its cue from an IDE. Chemistry borrows from the laboratory.

Subject-specific tools. Different subjects need different interactive instruments, and aime activates only those relevant to the lesson at hand — formula editor, graph plotter and geometry tools for Mathematics; code editor, terminal and live execution for Programming; molecular viewer, periodic table and reaction balancing for Chemistry; maps, globe viewer and timelines for Geography; diagram labelling and anatomy explorer for Biology. The full tool set, the activation rules and the way tools are bound to lesson intent sit inside aime's proprietary configuration.

The lesson pipeline

Every lesson passes through the same shape of pipeline before it reaches a learner: the subject selects its EduRule; ThinkBook supplies the pedagogical structure for the topic; the Planner Agent assembles the lesson against those constraints; the AIME Renderer delivers it inside a subject-specific environment with the subject's own theme, tools, whiteboard and slides. The orchestration that runs this pipeline — including the routing decisions, the composition strategy across models and the reconciliation rules at the edges — is handled by aimeFUSION inside the aime Loom™ Workflow Engine™.

“Most AI in education optimises what is taught. We optimise what is taught and how it is taught. The best way to teach Mathematics is not the best way to teach Biology, and the best way to teach Programming is not the best way to teach History. AI should adapt to the subject — not the other way around.”

Why this matters

• Subject-true pedagogy — every discipline is taught in its own idiom, not flattened into a single tutor voice.
• Cleaner cognitive surface — only the tools relevant to the lesson appear, so the environment helps the learner instead of crowding them.
• Live-classroom feel — the AIME Renderer mixes slides, annotation and freehand teaching the way a human teacher does.
• Composable, not monolithic — EduRule, ThinkBook and the renderer are independent components, which is what lets the system scale across subjects without rewriting the tutor for each one.

Availability

Lesson-specific learning environments are rolling out across the aime stack from June 2026, with the first wave covering Mathematics, Programming, Chemistry, Geography, Biology and Physics. Additional subjects, language coverage and age-stage variants follow through the second half of the year, distributed through aimeCLOUD™ for connected deployments and as self-contained .aimepack™ lessons for offline classrooms.

About aime

aime builds the operating system for educational intelligence — the foundational infrastructure layer that future education systems will run on. aime's stack combines structured curriculum knowledge, pedagogy-aware reasoning, compact education-tuned models, agentic orchestration and offline-capable deployment, and is designed for ministries of education, universities and national school systems.