The AI Taxonomy.
Terms  /  Architecture  /  System-Centric AI
05 · Architecture

System-Centric AI

An architecture where planning, state, memory, and control live outside the model.

System-Centric AI treats the model as a reasoning engine — powerful but stateless. Everything else — what to do next, what happened before, what's been decided, what tools are available — is managed by the system layer. The model is called when reasoning is needed and given exactly the context required for that specific reasoning task. This architecture separates concerns: the model reasons, the system remembers, plans, and orchestrates. The result is more predictable behavior, lower token cost, and the ability to swap models without losing accumulated knowledge.

Example
A coding agent built on System-Centric AI architecture works like this: the planning layer breaks a feature into tasks, the memory layer provides relevant codebase context for each task, the orchestration layer calls the model with a focused prompt for each task, and the verification layer checks the output before moving to the next task. The model never sees the full conversation history — it sees only what it needs for the current step. If the team switches from one model to another, the plans, memory, and verification logic remain intact.

More in Architecture

Model-Centric AI

An architecture where the model is expected to infer state, intent, plan, and memory from raw context.

Tool Isolation

Memory and context don't cross tool boundaries — each tool is a silo.

The Expertise Trap

Context and knowledge locked on one person's machine, inaccessible to the team.

Agent Fragmentation

Agents operating in disconnected silos with no shared state, duplicating work and contradicting each other.

Unified Context

A single, shared knowledge layer accessible to all agents and tools. The cure for Tool Isolation.

Context Portability

The ability to move knowledge between tools, agents, sessions, and team members without loss. The cure for The Expertise Trap.

Stateful, Inspectable Memory

A collaborative memory space where multiple agents and humans contribute to and draw from the same knowledge base — with full history and traceability.

Context Handoff

The ability to transfer full working context from one agent, session, or tool to another without loss. A specific operation enabled by Context Portability.