Architecture Map — Every Keyword's Place and Role in One View
LLM, Agent, Tool Calling, MCP, Skills, Workflow, Memory, Knowledge Graph, GraphRAG, and RAG are not parallel concepts of equal rank. Each has its own post.
About This Document
In enterprise AI adoption, a crowd of keywords tends to be presented as "options" at the same level. In reality, they are parts with different roles, arranged in layers with dependencies. This page provides that overall view, and doubles as a signpost to the section of this site where each keyword is covered.
Audience: First-time readers of this site; developers and architects who want the big picture of AI adoption
1. How the Terms Relate
| Term | In One Phrase | Contrasting Concept |
|---|---|---|
| LLM | A function that predicts | Agent (a subject that loops) |
| Agent | Autonomous judgment loop | Workflow (fixed procedure) |
| Tool Calling | The calling mechanism | MCP (its standard protocol) |
| Memory | What to remember | Context Window (volatile) |
| Knowledge Graph | Structured relations | Vector DB (similarity) |
| GraphRAG | Retrieval that walks relations | Plain RAG (fragment retrieval) |
2. Resource Types × Access Paths
Most of these keywords can be organized as pairs of "which resource, accessed by which means". In this diagram, Agent / Workflow sit on the orchestration side. In the II.1 five layers, however, Agent is one ownership post — not the top of a resource stack.
| Resource Type | Nature | Access Path | Notes |
|---|---|---|---|
| Document knowledge | Unstructured, static | RAG | Read-only "search by meaning". Extends to GraphRAG for cross-document questions |
| Business data | Structured, dynamic | DB (SQL / Semantic Layer) | Read access for "exact values". The LLM only generates the query |
| Business operations | Side effects | API | Write and execute — includes irreversible operations, so permission design is mandatory |
| Relational knowledge | Graph-structured | Knowledge Graph / Memory | "Who owns what, what depends on what". Retrieved via GraphRAG |
| Multi-step execution | Combination of the above | Agent / Workflow | The orchestration layer spanning resources. Fixed steps → Workflow; judgment needed → Agent |
IMPORTANT
The value of this classification is that "read vs write" separates naturally. RAG and DB are reads (safe, idempotent); API is operations (side effects, permissions required). When you grant an Agent permissions, this boundary becomes the risk boundary. See Permission vs Authority.
NOTE
MCP does not get its own row in this table — it runs across it, as the connection standard unifying RAG / DB / API access. Skills are the static knowledge and procedures the Agent layer consults; they belong to "defining the Agent's behavior", not to access paths.
3. Data Flow — a Cycle, Not a One-Way Street
The two return flows are the point. The Business Process generates new Data, and the Agent's execution experience accumulates in Memory, growing the Knowledge. Without these return flows, you fall back into the scatter-gather problem of "re-researching everything from scratch every time".
WARNING
If your data is fundamentally scattered, AI is not the solution. Put RAG or an Agent on top of scattered, dirty data and it will only reproduce the scatter faster. Data curation comes first — unifying relations in a Knowledge Graph, unifying metric definitions in a Semantic Layer.
4. Relation to the Five Layers
The diagrams on this page are a map of resources, access paths, and circulation. The main map of this book is the II.1 five layers. The axes differ. Do not read this page as another name for the five layers.
| Five layers | How it appears on this page |
|---|---|
| Doctrine | Not a node in the diagrams. The measure of purpose, prohibitions, and priority. See the table below and III.3 Doctrine |
| Agent | On the orchestration side. In the five layers it owns understanding and assignment — it is not an upper stack that owns the other layers |
| Skills | Not an access path. Stable knowledge and procedures the Agent reads. Kept off the diagram |
| Memory | Relational knowledge, and one end of the feedback from runs |
| MCP | The connection standard that cuts across RAG / DB / API |
The execution boundary (Harness / Hooks) is none of the five layers. It is a machine interrupt at a point in the run — stop, record, after-steps. It does not add a layer.
NOTE
When unsure, decide who owns it with the five layers first. Then use this page for which resource, read or write. Reverse the order and ownership mixes with connection.
5. Keyword → Section of This Site
| Keyword | Covered In | Role |
|---|---|---|
| LLM (structural constraints) | I.1 Constraint summary / Sister site: understanding-llm | Why and premises |
| Ownership split (five layers) | II.1 Five layers / II.2 Placement | Main map of this book |
| Doctrine (decision criteria) | III.3 Doctrine | Purpose, prohibitions, priority |
| Skills | III.1 Skills | Static knowledge and procedures |
| MCP / Tool Calling | III.2 MCP | Connection as an implementation mechanism |
| Memory / Knowledge Graph | III.4 Memory | How memory and relations persist |
| Agent / Sub-agent / A2A | III.5 Agent | Taxonomy and design of executors |
| RAG / GraphRAG | This page §2 (no dedicated page) | Read access to document knowledge. For types, see IV.1 Patterns |
| Semantic Layer | MCP / Semantic Layer | Design discipline for structured data access |
| Workflow | Workflows | Patterns for fixed procedures |
| Patterns / limits | IV.1 Patterns / IV.2 Limits | Choosing a type and how far it reaches |
| Permission / Authority | Permission vs. Authority | Separating permission from authority |
| Hooks | Hooks | Harness-side execution boundary. Not a layer |
TIP
When unsure which means to pick, three axes decide it: freshness (static → RAG, dynamic → DB/API), amount of judgment (none → Workflow, much → Agent), and state of the data (dirty → curate first; AI comes last).
Related Documents
- Overview (Information) — Positioning and structure of this section
- II.1 Five layers — Main map of this book (ownership)
- II.2 Placement — Where to put what
- Hooks — Execution boundary, not a layer
- IV.1 Patterns — Which type to pick when
Going Deeper: Why LLMs Need an External Information Foundation
This page covered the structure (What/How) of information architecture. To understand why an LLM alone is not enough — from the LLM's structural constraints — see the sister site.
- understanding-llm (top page) — The eight structural constraints (Context Rot, Knowledge Boundary, etc.) that make external references necessary
Last updated: August 2026