AI Education and Business AI Systems — The Complete PAIR Product Landscape

What We're Building

The arrival of practical AI raises two fundamental questions:

The first is education. What should children actually learn, and how? Traditional programming education teaches "writing code," but the core competency of the future is not manual coding — it is the ability to direct AI tools to create real things.

The second is business. Every day, companies switch between ERP, email, approvals, and communication systems. Enormous amounts of repetitive work drain human energy. PAIR's approach is simple: deploy AI as a 24/7 digital worker that operates these systems, so humans only need to make decisions.

---

Part One: AI & Robotics Education Programme

Our Educational Philosophy

• AI is not an elective — it is the new literacy. Just as computers and the internet became essential skills, AI is becoming a foundational capability everyone needs. We're not training programmers — we're training people who can use AI to create and solve problems.

• Real learning requires visible outcomes. Every course is designed so students produce something tangible, accessible, and genuinely theirs — not a practice file, but a real project live on the internet.

• Age-tiered, but never ability-capped. Our four course tracks provide clear progression paths, but exceptional students can jump ahead at any time. We offer flexible, combinable modules rather than a fixed pipeline.

• Industrial-grade tools, not "educational editions." Students use the same Codex that professional developers use, the same LeRobot framework that robotics researchers worldwide use. Real tools build real capability.

The Four Course Tracks

Track 1: AI-Powered Coding (Grade 5+)

Students use OpenAI Codex to build real web projects by directing AI in natural language. The course runs in three stages:

1. Understanding AI coding — Building the foundational insight: "the clearer you describe what you want, the better the AI delivers"
2. Building real projects — Project-based learning with a template library spanning beginner (personal homepage, image gallery) to advanced (AI chat apps, data dashboards), all web-based
3. Automated deployment & publishing — Every student gets their own subdomain (e.g., alex.ai-study.cn). One-click deployment. Projects live permanently online.

The core takeaway is not "coding ability" — it's AI collaboration skills, product thinking, and a digital identity.

Track 2: Early Robotics Discovery (Below Grade 5)

Block-based programming robots for young children. No AI concepts at all — the goal is planting the seed of computational thinking: "I can arrange instructions to make something move the way I want."

Track 3: Imitation Learning Robotics

This is the most revolutionary track. Traditional robotics courses follow a "human writes code → robot executes" model. Imitation learning flips this entirely: human demonstrates → robot observes → robot learns on its own.

Technology stack:

• LeRobot framework — HuggingFace's open-source robot learning platform
• VLA foundation models (π₀ / GR00T) — Vision-Language-Action models that see, understand language, and generate actions
• Fine-tuning — Using our recorded demonstration data to teach the model specific tasks on top of its pre-trained capabilities

Students transform from "programmers" into "coaches," moving through the complete loop: record demonstrations → cloud fine-tuning → deploy and verify → iterate and improve.

Track 4: Reinforcement Learning Robotics (Middle School+)

The most advanced track, built around Sim-to-Real — training in high-speed virtual environments first, then transferring learned strategies to physical robots.

Technology stack uses real industry tools: ROS 2 (the standard robot operating system) + NVIDIA Isaac Sim (high-fidelity simulation). Students progress through four phases: ROS 2 fundamentals → simulation environment setup → RL training → Sim-to-Real transfer.

Supporting Platforms

• Student AI Assistant — Each student gets their own cloud-hosted AI agent with persistent memory that understands their learning progress. Available 24/7. Monthly subscription.
• Moodle Learning Platform — The world's most widely used open-source LMS, hosting all structured course content.

---

Part Two: Pair Business AI System

Pair is a complete autonomous AI operating system for business. Its five-layer architecture:

Layer 1: White-Glove Dedicated Deployment

Each client gets an entirely independent Pair instance — dedicated server, dedicated Odoo database, dedicated AI engine. Complete data isolation.

Layer 2: Office Suite Integration

Full connectivity with Google Workspace and Microsoft 365 — Gmail, Drive, Docs, Calendar, Outlook, Teams all accessible. AI can read and write cloud documents directly.

Layer 3: Business Systems — Odoo Open-Source ERP

Odoo covers finance, sales, purchasing, inventory, HR, and every core business module. Its robust RPC (Remote Procedure Call) interface lets AI operate the ERP programmatically — fast, accurate, and verifiable.

Layer 4: AI Engine — Hermes

Our open-source AI agent framework. Core capabilities include tool invocation, long-term memory, multi-model support, and event-driven execution. Open-source means no vendor lock-in.

Layer 5: Communication & Interaction

Unified messaging across WeChat, Feishu (Lark), email, and Teams. AI serves as the central processing node — no information gets lost, everything gets a response.

Business Model

Three-tier pricing:

• Deployment fee (one-time) — Server provisioning, system installation, data migration, training
• Subscription (monthly/annual) — Ongoing operations, model upgrades, technical support
• Custom development — Industry-specific modules, legacy system integration, specialised automation

---

Partnership Opportunities

We're actively seeking two types of partners:

• Education channel partners — Organisations with access to schools or training centres
• Enterprise integration partners — IT service companies and software integrators

Learn more at www.pairforce.ai

---

Based on PAIR product strategy discussions, June 2026